Trabajo Final - Visión por computadora 2 - CEIA ¶
📝 Nota:
La funcionalidad de visualización de jupyter notebooks en github es solamente un preview.
Para mejor visualización se sugiere utilizar el visualizador recomndado por la comunidad: nbviewer
Puedes a acceder al sigiente enlace para ver este notebook en dicha página: TrabajoFinalVPC2
Figura 1: Detección de palmeras mediante modelos de aprendizaje profundo.
✨Datos del proyecto:✨
| Subtitulo | Trabajo final - VPC2 - FIUBA |
|---|---|
| Descrpción | Palm tree detector |
| Integrantes | - Bruno Masoller (brunomaso1@gmail.com) - Simón Rodriguez (simon.andre.r@gmail.com) |
✋ Importaciones: ✋
Inicialmente, se descargan varios datos desde Google Drive. Los archivos descargados constan de las carpetas datasets y utils. En la carpeta datasets se encuentran los conjuntos de datos y en la carpeta utils, utilidades necesarias, como archivos auxiliares que facilitan la visualización y entrenamiento del modelo.
In [ ]:
# Librerías básicas para el tratamiento de archivos del sistema y procesamiento aleatorio.
import sys, os, warnings, zipfile, shutil ,random, time, json, gc
# Configuraciónes previas.
DOWNLOAD_RESOURCES = True # Permite indicar si descargar el conjunto o no
RUTA_CARPETA_RESOURCES = './resources'
RUTA_ARCHIVO_REQUERIMIENTOS = os.path.join(RUTA_CARPETA_RESOURCES, 'requirements.txt')
RUTA_CARPETA_UTILS = os.path.join(RUTA_CARPETA_RESOURCES, 'utils')
# Esta lógica descarga los recursos necesarios para este proyecto desde Google Drive.
# Si no está disponible el link, se puede acceder desde:
# - Dataset Palmeras: https://www.kaggle.com/datasets/riotulab/aerial-images-of-palm-trees
# - Utilidades de torchvision: https://github.com/pytorch/vision/tree/main/references/detection
if DOWNLOAD_RESOURCES:
print("Descargando recursos...")
%pip install gdown
!gdown https://drive.google.com/drive/folders/1DFfgoGQ8-zLwQU85Uyq0X72GX4IITInV?usp=sharing --folder
# Verificamos si la carpeta ha sido descargada correctamente
if os.path.exists(RUTA_CARPETA_RESOURCES):
print(f"La carpeta {RUTA_CARPETA_RESOURCES} fue encontrada.")
# Verificamos si el archivo de requerimientos existe
if os.path.exists(RUTA_ARCHIVO_REQUERIMIENTOS):
print("Archivo de requerimientos encontrado, instalando dependencias...")
%pip install -r {RUTA_ARCHIVO_REQUERIMIENTOS}
else:
warnings.warn(f"El archivo {RUTA_ARCHIVO_REQUERIMIENTOS} no fue encontrado. No se instalarán los requerimientos.")
if os.path.exists(RUTA_CARPETA_UTILS):
# Se agrega esta ruta al path para que funcionen las utilidades de pytorch.
sys.path.append(os.path.abspath('./resources/utils'))
print(f"Se agregó la ruta {RUTA_CARPETA_UTILS} al system path.")
else:
raise FileNotFoundError(f"La carpeta {RUTA_CARPETA_UTILS} no fue encontrada. Verifique la descarga.")
else:
raise FileNotFoundError(f"La carpeta {RUTA_CARPETA_RESOURCES} no fue encontrada. Verifique la descarga.")
# Librería que muestra utilización de la GPU.
from GPUtil import showUtilization as gpu_usage
# Librería para TypesHint.
from typing import List, Tuple, Dict, Union, Callable
from enum import Enum
# Librería para el procesamiento numérico.
import numpy as np
# Librería para procesamiento de gráficas.
import matplotlib.pyplot as plt
import matplotlib.patches as patches
# librería de manejo de datos
import pandas as pd
# Librería de deeplearning de pytorch y su análogo para visión por computadora.
import torch, torchvision
from torch import nn, optim
# Utilidades de pytorch.
from torch.utils.data import DataLoader
from torch.utils.data import Dataset
# Utilidades de torchvision.
from torchvision.transforms import functional as F
from torchvision.transforms import v2 as T
from torchvision.ops import nms
from torchvision.io import read_image
from torchvision.utils import draw_bounding_boxes
from torchvision import tv_tensors
# Librería de aumentacion de datos
import albumentations as A
from albumentations.pytorch import ToTensorV2
# Librería de barritas de progreso
from tqdm.auto import tqdm
from tqdm import trange
# Librería para procesamiento de archivos XML.
import xml.etree.ElementTree as ET
# Librería de utilizadades para el procesamiento de visión por computadora.
import cv2
# Libreria para tratamiento de imágenes
# from PIL import Image
# Librería para visualiación de Datasets.
import fiftyone as fo
# Librería que permite ver el resumen de un modelo de deeplearning.
from torchinfo import summary
import ultralytics
from ultralytics import YOLO
# Librería de utilidades de pytorchvision/detection.
from engine import train_one_epoch, evaluate # type: ignore
from utils import MetricLogger
from coco_eval import CocoEvaluator
ultralytics.checks()
🔧 Configuraciones: 🔧
In [3]:
VERBOSE = True # Muestra época a época la evolución
LOAD_MODEL_FROM_CHECKPOINT = True # Permite cargar el modelo desde un checkpoint.
DEVICE = 'cuda' if torch.cuda.is_available() else 'cpu' # Establece el dispositivo.
RANDOM_SEED = 42
EPOCHS = 100 # Cantidad de épocas genéricas a entrenar.
BATCH_SIZE = 64 # Tamaño de batch genérico.
# Datasets:
PALMERAS_KAGGLE_DATSET = 'Palm-Counting-349images'
# Formatos
PALMERAS_KAGGLE_DATASET_VOC = 'palmeras_kaggle_voc_format'
PALMERAS_KAGGLE_DATASET_COCO = 'palmeras_kaggle_coco_format'
PALMERAS_KAGGLE_DATASET_YOLO = 'palmeras_kaggle_yolo_format'
# Rutas genéricas:
RUTA_CARPETA_COMPRESSED = os.path.join(RUTA_CARPETA_RESOURCES, 'datasets', 'compressed')
RUTA_CARPETA_RAW = os.path.join(RUTA_CARPETA_RESOURCES, 'datasets', 'raw')
RUTA_CARPETA_MODELS = os.path.join(RUTA_CARPETA_RESOURCES, 'models')
RUTA_CARPETA_KAGGLE_DATASET = os.path.join(RUTA_CARPETA_RAW, PALMERAS_KAGGLE_DATSET)
# Rutas dataset formato VOC
RUTA_CARPETA_KAGGLE_DATASET_VOC = os.path.join(RUTA_CARPETA_RAW, PALMERAS_KAGGLE_DATASET_VOC)
RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC, "train")
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC, "test")
RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN, "data")
RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_LABELS = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN, "labels")
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST, "data")
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST, "labels")
# Rutas datset formato yolo
RUTA_CARPETA_KAGGLE_DATASET_YOLO = os.path.join(RUTA_CARPETA_RAW, PALMERAS_KAGGLE_DATASET_YOLO)
RUTA_CARPETA_KAGGLE_DATASET_YOLO_TRAIN = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_YOLO, 'train')
RUTA_CARPETA_KAGGLE_DATASET_YOLO_TEST = os.path.join(RUTA_CARPETA_KAGGLE_DATASET_YOLO, 'test')
RUTA_CARPETA_KAGGLE_DATSET_YOLO_TRAIN_CSV = os.path.join(RUTA_CARPETA_KAGGLE_DATASET, 'train_labels.csv')
RUTA_CARPETA_KAGGLE_DATSET_YOLO_TEST_CSV = os.path.join(RUTA_CARPETA_KAGGLE_DATASET, 'test_labels.csv')
# Configuraciones de imágenes:
FORMATOS_IMAGENES = ('.jpg', '.jpeg', '.png')
FORMATOS_LABELS = ('.xml')
IMG_SIZE = 300
# Etiquetas:
LABELS = {
'Palm' : 1
}
# Modelos
FASTERRCNN_RESNET50_FPN_V2_MODEL_ID = 'fasterrcnn_resnet50_fpn_v2'
FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG = 'fasterrcnn_resnet50_fpn_v2_sin_aug'
YOLOV8_MODEL_ID = 'yolov8'
YOLOV11_MODEL_ID = 'yolov11'
YOLOV8_CONFIG_YAML = os.path.join(RUTA_CARPETA_UTILS, YOLOV8_MODEL_ID + '.yaml')
TRAIN_TEST_SPLIT_RATIO = [0.8, 0.2]
random.seed(RANDOM_SEED)
def clean_gpu_usage() -> None:
"""Permite mostrar el uso y limpiar el chache de la GPU"""
print("Initial GPU Usage")
gpu_usage()
gc.collect()
torch.cuda.empty_cache()
print("GPU Usage after emptying the cache")
gpu_usage()
print(f'Dispositivo actual: {DEVICE}')
Dispositivo actual: cpu
Introduction¶
En el marco de la especialización en Inteligencia Artificial de la Facultad de Ingeniería de la Universidad de Buenos Aires, se plantea, como trabajo final de la materia Visión por Computadora 2, se tiene como objetivo desarrollar un modelo de visión por computadora que utilice aprendizaje profundo para detectar ubicaciones de palmeras en diversas imágenes áereas. Metodológicamente, se aplican principios de CRISP-ML(Q) para el desarrollo del problema.
Business and data understanding¶
La detección de palmeras es ampliamente estudiado en el área de visión por computadora, dado su potencial como insumo de otras soluciones, como por ejemplo, aquellas que intentan mejorar la salud de las palmeras, donde inicalmente es deseable una detección aera para luego realizar una inspección manual (con el objetivo de reducir costos). Muchos estudios (como este por ejemplo) se realizan en paises en donde la mayoría de las palmeras son materia prima para la producción del "aceite de palmera". Sin embargo, en paises no productores, el cuidado de las palmeras también es importante, dado su impacto en el medio ambiente. El objetivo de este notebook es presentar enfoques de modelos que permitan detectar palmeras en diferentes conjuntos de datos.
Data preparation¶
Los datos fueron obenidos del sito de Kaggle: https://www.kaggle.com/datasets/riotulab/aerial-images-of-palm-trees, sin embargo, para una mejor manejo, se los importan desde Google Drive (https://drive.google.com/drive/folders/1DFfgoGQ8-zLwQU85Uyq0X72GX4IITInV).
Data pre-processing¶
Luego de descargado los datos, tenemos la siguiente estructura de directorios:
resources
├── datasets
│ └── compressed
│ └── palmeras_kaggle.zip
└── utils
├── engine.py
├── train.py
└── ...
🔮 Función auxiliar:
Dada una carpeta de origen y una carpeta de destino, descomprime todos los datos de la carpeta de origen en la carpeta de destino
In [17]:
def descomprimir_archivos(carpeta_origen: str, carpeta_destino: str) -> None:
"""
Dada una carpeta de origen y una carpeta de destino, descomprime todos los datos de la carpeta de origen en la carpeta de destino.
Args:
carpeta_origen (str): Ruta de la carpeta que contiene los archivos zip.
carpeta_destino (str): Ruta de la carpeta donde se descomprimirán los archivos.
Returns:
None
"""
archivos: List[str] = [f for f in os.listdir(carpeta_origen) if f.endswith(".zip")]
if not archivos:
warnings.warn("No se encontraron archivos .zip para descomprimir.")
return
for archivo in archivos:
ruta_archivo: str = os.path.join(carpeta_origen, archivo)
print(f"Descomprimiendo {archivo} en {carpeta_destino}...")
# Descomprimir archivo
with zipfile.ZipFile(ruta_archivo, "r") as zip_ref:
zip_ref.extractall(carpeta_destino)
print(f"{archivo} descomprimido exitosamente.")
Descomprimimos los datos:
In [18]:
# Crear la carpeta 'raw' si no existe
os.makedirs(RUTA_CARPETA_RAW, exist_ok=True)
# Verificar si la carpeta compressed existe
if os.path.exists(RUTA_CARPETA_COMPRESSED):
descomprimir_archivos(RUTA_CARPETA_COMPRESSED, RUTA_CARPETA_RAW)
else:
warnings.warn(f"La carpeta {RUTA_CARPETA_COMPRESSED} no fue encontrada. Se descargaron correctamente los datos?")
Descomprimiendo palmeras_kaggle.zip en ./resources\datasets\raw... palmeras_kaggle.zip descomprimido exitosamente.
Una vez descomprimido los archivos, podemos observar que el conjunto está en formato PascalVOC (aunque también tiene dos archivos .csv que pueden utilizarse para otros formatos, ej: YOLO).
Ajustamos los archvios para tener una carpeta "data" y otra "labels", de la siguiente forma:
palmeras_kaggle_formato_voc/
├── test/
│ ├── data/
│ │ ├── img1.jpg
│ │ ├── img2.jpg
│ │ └── ...
│ └── labels/
│ ├── img1.xml
│ ├── img2.xml
│ └── ...
└── train/
└── ...
🔮 Función auxiliar:
Mueve archivos de imágenes y etiquetas a sus respectivos directorios de destino.
In [19]:
def mover_archivos(source_dir: str, data_dest: str, label_dest: str) -> None:
"""
Mueve archivos de imágenes y etiquetas a sus respectivos directorios de destino.
Args:
source_dir (str): Directorio fuente que contiene los archivos a mover.
data_dest (str): Directorio de destino para los archivos de imágenes.
label_dest (str): Directorio de destino para los archivos de etiquetas.
Returns:
None
"""
for filename in os.listdir(source_dir):
file_path: str = os.path.join(source_dir, filename)
# Verifica si es un archivo de imagen según los formatos establecidos.
if filename.endswith(FORMATOS_IMAGENES):
shutil.move(file_path, os.path.join(data_dest, filename))
# Verifica si es un archivo de labels (xml comúnmente)
elif filename.endswith(FORMATOS_LABELS):
shutil.move(file_path, os.path.join(label_dest, filename))
🔮 Función auxiliar:
Parsea el dataset descomprimido a la estructura Pascal VOC
In [20]:
def parse_to_pascal_voc(desompresed_folder: str) -> None:
"""Parsea el dataset descomprimido a la estructura Pascal VOC"""
# Define los paths de las carpetas
base_dir = os.path.join(RUTA_CARPETA_RAW, desompresed_folder)
train_dir = os.path.join(base_dir, "train")
test_dir = os.path.join(base_dir, "test")
if not os.path.exists(base_dir):
warnings.warn(
f"No se descomprimió el archivo? No existe la carpeta {base_dir}."
)
return None
if os.path.exists(RUTA_CARPETA_KAGGLE_DATASET_VOC):
warnings.warn(
f"Ya existe la carpeta {RUTA_CARPETA_KAGGLE_DATASET_VOC}, no se hizo nada."
)
return None
# Crea las nuevas carpetas si no existen
os.makedirs(RUTA_CARPETA_KAGGLE_DATASET_VOC, exist_ok=True)
os.makedirs(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, exist_ok=True)
os.makedirs(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS, exist_ok=True)
os.makedirs(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA, exist_ok=True)
os.makedirs(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_LABELS, exist_ok=True)
# Mover archivos de train y test a las nuevas carpetas
try:
mover_archivos(
train_dir,
RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA,
RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_LABELS,
)
mover_archivos(
test_dir,
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA,
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS,
)
except:
warnings.warn(
f"Error al mover los archivos. Las carpetas {train_dir} y {test_dir} existen? O ya fueron borradas?"
)
Para ajustar a la estructura necesaria, realizamos el siguiente procesamiento:
In [21]:
parse_to_pascal_voc(PALMERAS_KAGGLE_DATSET)
🔮 Función auxiliar:
Cuenta la cantidad de archivos de imágenes que tiene un directorio
In [22]:
def contar_imagenes(directorio: str) -> int:
"""Cuenta la cantidad de archivos de imágenes que tiene un directorio"""
return len([f for f in os.listdir(directorio) if f.endswith(FORMATOS_IMAGENES)])
Luego que tenemos la estructura correcta, podemos visualizar la cantidad de ejemplos que tenemos:
In [23]:
print(f'Total de imágenes test: {contar_imagenes(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA)}')
print(f'Total de imágenes train: {contar_imagenes(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA)}')
Total de imágenes test: 70 Total de imágenes train: 279
Si inspeccionamos algunas imágenes, vemos que tienen un tag polygon en el xml (se utilizan usualmente en problemas de segmentación). También el tag path contiene una ruta que no corresponde. Para procesar estos archivos correctamente, se eliminan dichos tags
🔮 Función auxiliar:
Modifica un archivo XML eliminando objetos con etiquetas de polygon y 'path'.
In [24]:
def process_tags(xml_path: str) -> Tuple:
"""
Modifica un archivo XML eliminando objetos con etiquetas de polygon y 'path'.
Esta función analiza un archivo XML, elimina todos los objetos que contienen
una etiqueta polygon y path si existen. Los cambios se guardan en el mismo archivo.
Args:
xml_path (str): Ruta al archivo XML que se va a modificar.
Returns:
None
"""
# TODO: Falta cambiar el folder y ajustar el nombre del archivo.
tree: ET.ElementTree = ET.parse(xml_path)
root: ET.Element = tree.getroot()
tag_path_eliminados = 0
tag_polygon_eliminados = 0
# Eliminar el tag 'path' si existe
path_elem = root.find("path")
if path_elem is not None:
root.remove(path_elem)
tag_path_eliminados += 1
# Encontrar todos los objetos
for obj in root.findall("object"):
# Si el objeto tiene un tag polygon, eliminar el objeto
if obj.find("polygon") is not None:
root.remove(obj)
tag_polygon_eliminados += 1
# Guardar los cambios en el archivo
tree.write(xml_path)
return tag_path_eliminados, tag_polygon_eliminados
🔮 Función auxiliar:
Recorre todos los archivos xml y aplica process_tags
In [25]:
def apply_process_tags(base_dir: str, formatos_labels: Tuple):
total_tag_path_eliminados = 0
total_tag_polygon_eliminados = 0
for xml_file in os.listdir(base_dir):
if xml_file.endswith(formatos_labels):
xml_file_path = os.path.join(base_dir, xml_file)
# Aplicar la función a cada archivo XML
tag_path_eliminados, tag_polygon_eliminados = process_tags(xml_file_path)
total_tag_path_eliminados += tag_path_eliminados
total_tag_polygon_eliminados += tag_polygon_eliminados
return total_tag_path_eliminados, total_tag_polygon_eliminados
Recorremos todos los archivos XML y eliminamos las etiquetas:
In [26]:
total_tag_path_eliminados = 0
total_tag_polygon_eliminados = 0
tag_path_eliminados, tag_polygon_eliminados = apply_process_tags(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS, FORMATOS_LABELS)
total_tag_path_eliminados += tag_path_eliminados
total_tag_polygon_eliminados += tag_polygon_eliminados
tag_path_eliminados, tag_polygon_eliminados = apply_process_tags(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_LABELS, FORMATOS_LABELS)
total_tag_path_eliminados += tag_path_eliminados
total_tag_polygon_eliminados += tag_polygon_eliminados
print(f"Tags polygons eliminados: {total_tag_polygon_eliminados}")
print(f"Tags path eliminados: {total_tag_path_eliminados}")
Tags polygons eliminados: 3 Tags path eliminados: 349
Data visualization¶
🔮 Función auxiliar:
Funciones auxiliares para visualización de las imágenes.
In [30]:
def paint_bounding_boxes(
ax: plt.Axes, boxes: np.ndarray, labels: list, config: Dict
) -> None:
"""
Dibuja bounding boxes en el eje especificado.
Args:
ax (plt.Axes): Eje de Matplotlib donde se dibujarán las cajas.
boxes (np.ndarray): Array de coordenadas de las cajas.
labels (list): Lista de etiquetas correspondientes a cada caja.
config (Dict): Configuración para el estilo de las cajas y texto.
Returns:
None
"""
for box, label in zip(boxes, labels):
xmin, ymin, xmax, ymax = box
rect = patches.Rectangle(
(xmin, ymin),
xmax - xmin,
ymax - ymin,
linewidth=config["linewidth"],
edgecolor=config["edgecolor"],
facecolor="none",
)
ax.add_patch(rect)
if config.get("putText", False):
ax.text(
xmin,
ymin - 10,
label,
color=config["color"],
fontsize=config["fontsize"],
backgroundcolor=config.get("backgroundcolor", "white"),
)
def draw_bbox(
img: Union[str, torch.Tensor],
target: Union[str, Dict[str, torch.Tensor]],
ax: plt.Axes,
config: Dict,
) -> None:
"""
Dibuja bounding boxes en una imagen desde un archivo o un dataset.
Args:
img (Union[str, torch.Tensor]): Ruta de la imagen o tensor de la imagen.
target_or_annotation_path (Union[str, Dict[str, torch.Tensor]]): Ruta al archivo XML de anotación o diccionario con anotaciones.
ax (plt.Axes): Eje de Matplotlib donde se dibujará la imagen.
config (Dict): Configuración para el estilo de las cajas y texto.
Returns:
None
"""
# Validación básica
if not isinstance(ax, plt.Axes):
raise ValueError("ax debe ser un objeto plt.Axes")
if isinstance(img, str):
image = read_image(img)
elif isinstance(img, torch.Tensor):
image = img
else:
raise ValueError("img debe ser una cadena o un tensor")
image_tensor = image.permute(1, 2, 0)
ax.imshow(image_tensor)
ax.axis("off")
if config.get("from_dataset_voc", False):
boxes = target["boxes"]
labels = target["labels"]
paint_bounding_boxes(ax, boxes, labels, config)
else:
tree = ET.parse(target)
root = tree.getroot()
boxes = []
labels = []
for obj in root.findall("object"):
bbox = obj.find("bndbox")
xmin = int(bbox.find("xmin").text)
ymin = int(bbox.find("ymin").text)
xmax = int(bbox.find("xmax").text)
ymax = int(bbox.find("ymax").text)
boxes.append((xmin, ymin, xmax, ymax))
labels.append(obj.find("name").text)
paint_bounding_boxes(ax, np.array(boxes), labels, config)
def draw_grid(
images: List[Union[str, torch.Tensor]],
targets: List[Union[str, Dict[str, torch.Tensor]]],
config: Dict,
grid_size: Tuple[int, int] = (3, 3),
figsize: Tuple[int, int] = (15, 15),
) -> None:
"""
Dibuja una grilla de imágenes con bounding boxes.
Args:
images (List[Union[str, torch.Tensor]]): Lista de rutas de imágenes o tensores de imágenes.
targets (List[Union[str, Dict[str, torch.Tensor]]]): Lista de rutas a archivos XML de anotación o diccionarios con anotaciones.
grid_size (Tuple[int, int], opcional): Tamaño de la grilla (filas, columnas). Por defecto es (3, 3).
from_dataset (bool, opcional): Indica si los datos provienen de un dataset. Por defecto es False.
Returns:
None
"""
fig, axs = plt.subplots(grid_size[0], grid_size[1], figsize=figsize)
for i in range(grid_size[0] * grid_size[1]):
img = images[i]
target = targets[i]
ax = axs[i // grid_size[1], i % grid_size[1]]
draw_bbox(img, target, ax, config)
# Mostrar el nombre de la imagen si está habilitado en config
img_name = (
os.path.basename(target["img_path"])
if config.get("from_dataset_voc", False)
else os.path.basename(img)
)
if config.get("show_image_name", False):
ax.annotate(
img_name,
xy=(0.5, -0.05),
xycoords="axes fraction",
ha="center",
fontsize=config.get("name_fontsize", 12),
)
# Agregar el super título si está definido en config
if config.get("super_title", ""):
fig.suptitle(
config["super_title"],
fontsize=config.get("super_title_fontsize", 16),
y=config.get("super_title_y", 0.95),
)
plt.tight_layout()
plt.show()
def show_single_image(
image: Union[str, torch.Tensor],
target: Union[str, Dict[str, torch.Tensor]],
config: Dict,
figsize: Tuple[int, int] = (8, 8),
) -> None:
"""
Muestra una sola imagen con sus bounding boxes.
Args:
image (Union[str, torch.Tensor]): Ruta de la imagen o tensor de la imagen.
target (Union[str, Dict[str, torch.Tensor]]): Ruta al archivo XML de anotación o diccionario con anotaciones.
from_dataset (bool, opcional): Indica si los datos provienen de un dataset. Por defecto es False.
Returns:
None
"""
fig, ax = plt.subplots(figsize=figsize)
draw_bbox(image, target, ax, config)
# Mostrar el nombre de la imagen si está habilitado en config
img_name = (
os.path.basename(target["img_path"])
if config.get("from_dataset_voc", False)
else os.path.basename(image)
)
if config.get("show_image_name", False):
ax.annotate(
img_name,
xy=(0.5, -0.05),
xycoords="axes fraction",
ha="center",
fontsize=config.get("name_fontsize", 12),
)
# Agregar el super título si está definido en config
if config.get("super_title", ""):
fig.suptitle(
config["super_title"],
fontsize=config.get("super_title_fontsize", 16),
y=config.get("super_title_y", 0.95),
)
plt.tight_layout()
plt.show()
def get_image_annotation_paths(
image_dir: str, label_dir: str, max_files: int = 9
) -> Tuple[List[str], List[str]]:
"""
Obtiene las rutas de imágenes y sus correspondientes anotaciones de forma aleatoria.
Args:
image_dir (str): Directorio que contiene las imágenes.
label_dir (str): Directorio que contiene las anotaciones XML.
max_files (int, opcional): Número máximo de archivos a procesar. Por defecto es 9.
Returns:
Tuple[List[str], List[str]]: Una tupla conteniendo dos listas:
- Lista de rutas de imágenes.
- Lista de rutas de anotaciones XML correspondientes.
"""
image_files: List[str] = [
f for f in os.listdir(image_dir) if f.endswith(FORMATOS_IMAGENES)
]
# Selecciona al azar hasta max_files imágenes
image_files = random.sample(image_files, min(len(image_files), max_files))
image_paths: List[str] = [os.path.join(image_dir, img) for img in image_files]
annotation_paths: List[str] = [
os.path.join(label_dir, os.path.splitext(img)[0] + ".xml")
for img in image_files
]
return image_paths, annotation_paths
def get_random_image_and_annotation(image_dir: str, label_dir: str) -> Tuple[str, str]:
"""
Obtiene una imagen aleatoria y su correspondiente anotación.
Args:
image_dir (str): Directorio que contiene las imágenes.
label_dir (str): Directorio que contiene las anotaciones XML.
Returns:
Tuple[str, str]: Una tupla conteniendo:
- Ruta de la imagen seleccionada aleatoriamente.
- Ruta de la anotación XML correspondiente.
"""
image_files: List[str] = [
f for f in os.listdir(image_dir) if f.endswith(FORMATOS_IMAGENES)
]
random_image: str = random.choice(image_files)
image_path: str = os.path.join(image_dir, random_image)
annotation_path: str = os.path.join(
label_dir, os.path.splitext(random_image)[0] + ".xml"
)
return image_path, annotation_path
Para iniciar nuestro análisis sobre el conjunto de imágenes, primeramente visualizamos una imagen con sus detecciones:
In [31]:
image_path, annotation_path = get_random_image_and_annotation(
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS
)
config = {
"linewidth": 2,
"edgecolor": "r",
"facecolor": "none",
"color": "white",
"fontsize": 8,
"backgroundcolor": "red",
"putText": True,
"show_image_name" : True,
"name_fontsize" : 10,
"super_title": "Visualización de un ejemplo de Palmeras anotadas",
"super_title_fontsize": 16,
"super_title_y": 0.90,
"from_dataset_voc" : False,
}
show_single_image(image_path, annotation_path, config)
Visualizamos también una grilla con varias imágenes y sus detecciones:
In [15]:
image_paths, annotation_paths = get_image_annotation_paths(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS, max_files=9)
config = {
"linewidth": 1,
"edgecolor": "r",
"facecolor": "none",
"color": "white",
"fontsize": 6,
"backgroundcolor": "red",
"putText": False,
"show_image_name" : True,
"name_fontsize" : 10,
"super_title": "Visualización de varios ejemplos de Palmeras anotadas",
"super_title_fontsize": 16,
"super_title_y": 0.95,
"from_dataset_voc" : False,
}
draw_grid(image_paths, annotation_paths, config)
También podemos utilizar la librería de FiftyOne (https://docs.voxel51.com/) para una mejor visualización. Otra ventaja de esta librería es que tiene un módulo de exportación a varios formatos conocidos:
In [16]:
# Cargamos el conjunto.
fiftyone_dataset_train = fo.Dataset.from_dir(
dataset_type=fo.types.VOCDetectionDataset,
dataset_dir=RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN,
name=PALMERAS_KAGGLE_DATASET_VOC + '_train',
overwrite=True
)
100% |█████████████████| 279/279 [11.3s elapsed, 0s remaining, 49.3 samples/s]
Podemos observar los datos del conjunto:
In [17]:
print(fiftyone_dataset_train)
Name: palmeras_kaggle_voc_format_train
Media type: image
Num samples: 279
Persistent: False
Tags: []
Sample fields:
id: fiftyone.core.fields.ObjectIdField
filepath: fiftyone.core.fields.StringField
tags: fiftyone.core.fields.ListField(fiftyone.core.fields.StringField)
metadata: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.metadata.ImageMetadata)
created_at: fiftyone.core.fields.DateTimeField
last_modified_at: fiftyone.core.fields.DateTimeField
ground_truth: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Detections)
También se puede lanzar la UI de la aplicacion:
session = fo.launch_app(fiftyone_dataset)
session.open_tab()
Figura 2: UI de FiftyOne para el dataset de las palmeras.
La ventaja de utilizar FiftyOne, es que se puede exportar a varios formatos admitidos (https://docs.voxel51.com/user_guide/export_datasets.html | https://docs.voxel51.com/user_guide/export_datasets.html#custom-dataset-exporter). Como ejemplo, para exportar al formato de COCO sería:
fiftyone_dataset.export(
export_dir=RUTA_CARPETA_KAGGLE_DATASET_COCO,
dataset_type=fo.types.COCODetectionDataset,
label_field="ground_truth",
)
Machine learning model engineering¶
Como experimentos, dado que inicialmente el modelo no tiene como objetivo el funcionamiento en tiempo real, se plantea utilizar modelos de detección en dos etapas (ej: Faster R-CNN). Sin embargo, también se tienen en cuenta modelos de una sola etapa (ej: YOLO)
Faster R-CNN¶
🤖 Consulta a IA (ChatGPT 4o): 🤖
❔ Prompt:
Podrías generar una introducción sobre el modelo Faster R-CNN? Que incluya sus componentes, sus ventajas y desventajas, así como también comparalo con otros modelos del estado del arte.
🔃 Respuesta:
Faster R-CNN (Region-based Convolutional Neural Networks) es un modelo de detección de objetos que destaca por su precisión y eficiencia, especialmente en tareas de detección de múltiples clases en imágenes. Es una evolución de los modelos R-CNN (2014) y Fast R-CNN (2015), integrando mejoras que reducen el tiempo de procesamiento mientras mantienen un alto rendimiento. Faster R-CNN se ha vuelto un estándar en aplicaciones de visión por computadora que requieren detección precisa de objetos, como en vehículos autónomos, vigilancia y reconocimiento de objetos a partir de imágenes aéreas o satelitales. Componentes de Faster R-CNN Red Convolucional Base: Faster R-CNN utiliza una red convolucional como base, comúnmente modelos preentrenados como VGG16, ResNet o MobileNet. Estas redes actúan como extractores de características que convierten una imagen de entrada en un mapa de características. Region Proposal Network (RPN): Uno de los principales avances de Faster R-CNN es la RPN. En lugar de generar propuestas de regiones usando métodos externos como Selective Search (usado en R-CNN y Fast R-CNN), la RPN genera las regiones propuestas (candidatas a contener objetos) directamente dentro de la red. La RPN es una pequeña red convolucional que desliza una ventana sobre el mapa de características, prediciendo posibles "anclas" que podrían contener objetos. RoI Pooling (Region of Interest Pooling): Después de generar las regiones propuestas, Faster R-CNN aplica una técnica llamada RoI Pooling para adaptar cada región propuesta a un tamaño fijo, lo que permite alimentar estas regiones a una red totalmente conectada (fully connected) para su clasificación y refinamiento de los bounding boxes. Clasificación y Refinamiento de Bounding Boxes: La última parte del modelo toma cada región propuesta y predice tanto la clase del objeto (o si es fondo) como las coordenadas refinadas del bounding box. Ventajas de Faster R-CNN Precisión: Faster R-CNN es uno de los modelos de detección de objetos más precisos, ya que optimiza tanto la propuesta de regiones como la detección de objetos en un solo paso. Eficiencia: La integración de la Region Proposal Network permite un proceso más eficiente comparado con métodos anteriores, ya que la red aprende a generar propuestas más relevantes sin depender de técnicas externas que incrementan el tiempo de procesamiento. Flexibilidad: Puede adaptarse a diferentes tareas de detección, desde objetos pequeños en imágenes detalladas hasta objetos más grandes en escenas menos densas. Desventajas de Faster R-CNN Velocidad: Aunque es significativamente más rápido que sus predecesores, sigue siendo más lento que modelos como YOLO o SSD, ya que sigue utilizando un enfoque de dos etapas (propuesta de regiones y clasificación). Complejidad Computacional: Requiere mayor poder computacional que modelos más livianos, lo que lo hace menos adecuado para aplicaciones en tiempo real o en dispositivos con restricciones de hardware. Comparación con otros Modelos del Estado del Arte YOLO (You Only Look Once): A diferencia de Faster R-CNN, YOLO es un modelo de detección de una sola etapa, lo que significa que realiza la detección de objetos en una sola pasada sobre la imagen. Esto lo hace extremadamente rápido y adecuado para aplicaciones en tiempo real, pero a menudo sacrifica algo de precisión en comparación con Faster R-CNN. YOLO predice directamente las clases y las coordenadas de los objetos desde la imagen completa, lo que es más eficiente pero puede resultar en detecciones menos precisas para objetos pequeños o con formas irregulares. SSD (Single Shot Multibox Detector): Al igual que YOLO, SSD es un modelo de detección de una sola etapa que es más rápido que Faster R-CNN. Sin embargo, SSD tiende a tener un rendimiento ligeramente inferior en la detección de objetos pequeños y es menos preciso cuando se compara con Faster R-CNN en tareas que requieren alta precisión. RetinaNet: Este modelo combina lo mejor de ambos mundos, con un enfoque de una etapa similar a YOLO y SSD, pero utilizando una pérdida focal que pone mayor peso en las clases raras o difíciles de detectar. Aunque es más rápido que Faster R-CNN, sigue siendo algo más lento que YOLO y SSD. Conclusión Faster R-CNN es un modelo robusto y preciso que sigue siendo uno de los más utilizados en tareas de detección de objetos donde la precisión es crítica. Aunque existen modelos más rápidos como YOLO y SSD, estos tienden a sacrificar precisión, particularmente en la detección de objetos pequeños. Por otro lado, modelos más recientes como RetinaNet buscan un equilibrio entre precisión y velocidad, pero Faster R-CNN sigue siendo la opción preferida cuando la prioridad es maximizar la exactitud de las detecciones.
Figura 3: Arquitectura de Faster R-CNN. Extraída de: https://towardsdatascience.com/faster-r-cnn-for-object-detection-a-technical-summary-474c5b857b46
Preparar el datset¶
En el formato actual del conjunto, es necesario generar una clase Dataset Customizada para el entrenamiento. Para eso podemos tomar como ejemplo este tutorial: https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html
📝 Nota:
Para entrenar la red Faster R-CNN, es necesario que se cumplan algunos requerimientos, entre los cuales se encuentra una definición de la estructura del Dataset como la siguiente: También destacar que el modelo considera a la clase 0 como el fondo, lo que implica consideración en las labels
boxes, torchvision.tv_tensors.BoundingBoxes of shape [N, 4]: the coordinates of the N bounding boxes in [x0, y0, x1, y1] format, ranging from 0 to W and 0 to Hlabels, integer torch.Tensor of shape [N]: the label for each bounding box. 0 represents always the background class.image_id, int: an image identifier. It should be unique between all the images in the dataset, and is used during evaluationarea, float torch.Tensor of shape [N]: the area of the bounding box. This is used during evaluation with the COCO metric, to separate the metric scores between small, medium and large boxes.iscrowd, uint8 torch.Tensor of shape [N]: instances with iscrowd=True will be ignored during evaluation.masks, torchvision.tv_tensors.Mask of shape [N, H, W]: the segmentation masks for each one of the objects
Primero creamos el conjunto personalizado sin las transformaciones:
📝 Nota:
En la creación de este conjunto, hay varias matices, entre las cuales se encuentra el tratamiento de ejemplos que no tienen ninguna detección.
In [18]:
class PalmTreeDataset(Dataset):
"""Clase que crea el dataset partiendo de las imágenes en las carpetas data y labels"""
def __init__(self, data_dir, labels_dir):
self.data_dir = data_dir
self.labels_dir = labels_dir
self.imgs = list(sorted(os.listdir(data_dir)))
self.labels = list(sorted(os.listdir(labels_dir)))
def __len__(self):
return len(self.imgs)
def __getitem__(self, idx):
# Cargar imagen
img_path = os.path.join(self.data_dir, self.imgs[idx])
label_path = os.path.join(self.labels_dir, self.labels[idx])
img = read_image(img_path)
# Parsear archivo XML (PascalVOC)
tree = ET.parse(label_path)
root = tree.getroot()
boxes = []
labels = []
isCrowd = []
for obj in root.findall("object"):
label = obj.find("name").text
if label in LABELS.keys():
bbox = obj.find("bndbox")
xmin = int(bbox.find("xmin").text)
ymin = int(bbox.find("ymin").text)
xmax = int(bbox.find("xmax").text)
ymax = int(bbox.find("ymax").text)
boxes.append([xmin, ymin, xmax, ymax])
labels.append(LABELS[label]) # Hay solo una clase: Palmera
isCrowd.append(0) # Seteamos como falso.
# Convertimos a tensores
# Tratamiento de cuando no hay detecciones.
if len(boxes) > 0:
boxes = tv_tensors.BoundingBoxes(
boxes, format="XYXY", canvas_size=F.get_image_size(img)
)
area = (boxes[:, 3] - boxes[:, 1]) * (boxes[:, 2] - boxes[:, 0])
else:
# No hay ningún objeto detectado.
boxes = torch.zeros((0, 4), dtype=torch.float32)
area = torch.tensor([], dtype=torch.float32)
# Convertir el nombre del archivo a un número basado en los caracteres
image_id = idx
isCrowd = torch.as_tensor(isCrowd, dtype=torch.int64)
labels = torch.as_tensor(labels, dtype=torch.int64)
img = tv_tensors.Image(img)
target = {}
target["boxes"] = boxes
target["labels"] = labels
target["image_id"] = image_id
target["area"] = area
target["iscrowd"] = isCrowd
target["img_path"] = img_path
return img, target
Definir las transformaciones¶
Si utilizaramos Albumentations, un posible pipeline de transofrmaciones sería:
# Utilizando Albumentations
def get_transforms(train=True):
transforms = []
if train:
transforms.append(A.ColorJitter(brightness=0.2, contrast=0.2, saturation=0.2, hue=0.1, p=0.5)) # Ajuste HSV
transforms.append(A.GaussianBlur(blur_limit=3, sigma_limit=(0.1, 2.0), p=0.5)) # Desenfoque gaussiano
transforms.append(A.Rotate(limit=10, p=0.5))
transforms.append(A.HorizontalFlip(p=0.5))
transforms.append(A.VerticalFlip(p=0.5))
transforms.append(A.Affine(scale=(0.8, 1.2), translate_percent=(0.1, 0.1), p=0.5))
transforms.append(A.LongestMaxSize(max_size=IMG_SIZE))
transforms.append(ToTensorV2())
return A.Compose(transforms, bbox_params=A.BboxParams(format='pascal_voc', label_fields=['labels']))
class TransformedPalmTreeDatasetAlbumentations(Dataset):
"""Clase de más alto rango que aplica las transformaciones al PalmTreeDataset"""
def __init__(self, dataset, transforms=None):
self.dataset = dataset
self.transforms = transforms
def __len__(self):
return len(self.dataset)
def __getitem__(self, idx):
img, target = self.dataset[idx]
# Convertir la imagen a formato NumPy para que funcione con Albumentations
img = img.permute(1, 2, 0).numpy() # Cambiar de CxHxW a HxWxC
boxes = target['boxes'].tolist()
labels = target['labels'].tolist()
if self.transforms is not None:
# Aplicar transformaciones
transformed = self.transforms(image=img, bboxes=boxes, labels=labels)
img = transformed['image']
boxes = transformed['bboxes']
labels = transformed['labels']
# Reconstruir los bounding boxes como un tensor
target['boxes'] = torch.tensor(boxes, dtype=torch.float32)
target['labels'] = torch.tensor(labels, dtype=torch.int64)
return img, target
Sin embargo, las transformaciones, si bien mejoran los resultados, como mostrado en el clásico paper de Google, también aumentan el procesamiento. En este punto, se definen estas simples transofrmaciones:
In [19]:
def get_transforms(train):
transforms = []
if train:
transforms.append(T.ColorJitter(brightness=0.2, contrast=0.2, saturation=0.2, hue=0.1))
transforms.append(T.GaussianBlur(kernel_size=3, sigma=(0.1, 2.0)))
transforms.append(T.Resize(size=IMG_SIZE))
transforms.append(T.ToDtype(torch.float, scale=True))
transforms.append(T.ToPureTensor())
return T.Compose(transforms)
Creamos un conjunto de más alto rango que aplique las transformaciones. Estas transformaciones se aplican de este modo para separar con anterioridad al conjunto de train/test.
In [20]:
class TransformedPalmTreeDataset(Dataset):
"""Clase de más alto rango que aplica las transofrmaciones al PalmsDataset"""
def __init__(self, dataset, transforms=None):
self.dataset = dataset
self.transforms = transforms
def __len__(self):
return len(self.dataset)
def __getitem__(self, idx):
img, target = self.dataset[idx]
if self.transforms is not None:
img, target = self.transforms(img, target)
return img, target
Creamos los datasets:
📝 Nota:
En este punto se podría dividir el conjuto en train/test/validation de la siguiente forma:
# Creamos el dataset
full_dataset = PalmTreeDataset(RUTA_CARPETA_KAGGLE_DATASET_VOC_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_LABELS)
# Lo dividimos en train/test
train_dataset, test_dataset = torch.utils.data.random_split(full_dataset, TRAIN_TEST_SPLIT_RATIO, torch.Generator().manual_seed(RANDOM_SEED))
Sin embargo, las imágnes ya están divididas, y para manetener la coherencia entre los modelos, se mantiene esta estrategia.
In [21]:
train_dataset_raw = PalmTreeDataset(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_LABELS)
test_dataset_raw = PalmTreeDataset(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS)
# Aplicamos las transformaciones.
train_dataset = TransformedPalmTreeDataset(train_dataset_raw, transforms=get_transforms(True))
test_dataset = TransformedPalmTreeDataset(test_dataset_raw, transforms=get_transforms(False))
Visualizamos las transformaciones para una imagen:
In [22]:
random_idx = torch.randint(0, len(train_dataset), (1,)).item()
image, target = train_dataset[random_idx]
output_image = draw_bounding_boxes(image, target['boxes'], colors='red', width=2)
plt.figure()
plt.imshow(output_image.permute(1, 2, 0))
Out[22]:
<matplotlib.image.AxesImage at 0x7e58404bc970>
O para una grilla de imágenes:
In [23]:
config = {
"linewidth": 2,
"edgecolor": "r",
"facecolor": "none",
"color": "white",
"fontsize": 6,
"backgroundcolor": "red",
"putText": False,
"show_image_name": True,
"name_fontsize": 10,
"super_title": "Visualización de data augmentation",
"super_title_fontsize": 16,
"super_title_y": 0.95,
"from_dataset_voc": True,
}
# Generar 9 índices aleatorios del dataset
random_indices = random.sample(range(len(train_dataset)), 9)
draw_grid(
[train_dataset[i][0] for i in random_indices],
[train_dataset[i][1] for i in random_indices],
config,
)
Definición del modelo¶
Utilizaremos el modelo pre-entrando en el dataset de COCO (https://pytorch.org/vision/stable/models/generated/torchvision.models.detection.fasterrcnn_resnet50_fpn_v2.html#torchvision.models.detection.fasterrcnn_resnet50_fpn_v2). Primero, observamos como está compuesto el modelo para cambiar las capas correspondientes para nuestro caso:
In [24]:
preview_model = torchvision.models.detection.fasterrcnn_resnet50_fpn_v2(weights=None)
print(preview_model)
FasterRCNN(
(transform): GeneralizedRCNNTransform(
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
Resize(min_size=(800,), max_size=1333, mode='bilinear')
)
(backbone): BackboneWithFPN(
(body): IntermediateLayerGetter(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): Bottleneck(
(conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer2): Sequential(
(0): Bottleneck(
(conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer3): Sequential(
(0): Bottleneck(
(conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(5): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer4): Sequential(
(0): Bottleneck(
(conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
)
(fpn): FeaturePyramidNetwork(
(inner_blocks): ModuleList(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(2048, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(layer_blocks): ModuleList(
(0-3): 4 x Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(extra_blocks): LastLevelMaxPool()
)
)
(rpn): RegionProposalNetwork(
(anchor_generator): AnchorGenerator()
(head): RPNHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): ReLU(inplace=True)
)
)
(cls_logits): Conv2d(256, 3, kernel_size=(1, 1), stride=(1, 1))
(bbox_pred): Conv2d(256, 12, kernel_size=(1, 1), stride=(1, 1))
)
)
(roi_heads): RoIHeads(
(box_roi_pool): MultiScaleRoIAlign(featmap_names=['0', '1', '2', '3'], output_size=(7, 7), sampling_ratio=2)
(box_head): FastRCNNConvFCHead(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(4): Flatten(start_dim=1, end_dim=-1)
(5): Linear(in_features=12544, out_features=1024, bias=True)
(6): ReLU(inplace=True)
)
(box_predictor): FastRCNNPredictor(
(cls_score): Linear(in_features=1024, out_features=91, bias=True)
(bbox_pred): Linear(in_features=1024, out_features=364, bias=True)
)
)
)
Como haermos fine-tuning de este modelo, lo único que debemos cambiar es las últmas capas:
(box_predictor): FastRCNNPredictor(
(cls_score): Linear(in_features=1024, out_features=91, bias=True)
(bbox_pred): Linear(in_features=1024, out_features=364, bias=True)
)
Definimos una función especial para esto, en donde se le pasa el número de clases que debe predecir. También permite obtener el modelo desde el útimo checkpoint de entrenamiento:
In [25]:
def get_fasterrcnn_resnet50_fpn_v2_model(num_classes: int, checkpoint_path: str = None, device: str = 'cpu') -> torch.nn.Module:
# Cargar un modelo preentrenado de Faster R-CNN (los pesos de coco)
model = torchvision.models.detection.fasterrcnn_resnet50_fpn_v2(weights='DEFAULT')
# Obtener el número de entradas de la capa de clasificación
in_features = model.roi_heads.box_predictor.cls_score.in_features
# Reemplazar la cabeza del clasificador por una que tenga el número de clases deseado
model.roi_heads.box_predictor = torchvision.models.detection.faster_rcnn.FastRCNNPredictor(in_features, num_classes + 1)
# Optimizer
params = [p for p in model.parameters() if p.requires_grad]
optimizer = optim.SGD(params, lr=0.005, momentum=0.9, weight_decay=0.0005)
# Aprender tasa de aprendizaje (LR) programada
lr_scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=3, gamma=0.1)
# Cargar los pesos del modelo desde un checkpoint si fue especificado.
model.to(device)
data = {}
if checkpoint_path and os.path.exists(checkpoint_path):
print(f"Cargando el modelo desde el checkpoint: {checkpoint_path}")
# Cargar los parámetros del checkpoint
checkpoint = torch.load(checkpoint_path, map_location=device)
# Restaurar el estado del modelo
model.load_state_dict(checkpoint['model_state_dict'])
optimizer.load_state_dict(checkpoint['optimizer_state_dict'])
lr_scheduler.load_state_dict(checkpoint['scheduler_state_dict'])
data = checkpoint['data']
print(f"\nModelo cargado desde el checkpoint en la época {data['epoch']}.\n")
else:
print("\nCargando el modelo desde 0.\n")
return model, optimizer, lr_scheduler, data
Finalmente, instanciamos el modelo y vemos si efectivamente se cambiaron las últimas capas:
In [26]:
# Instanciamos el modelo.
save_checkpoint_path = os.path.join(RUTA_CARPETA_MODELS, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID + '_checkpoint.tar')
if LOAD_MODEL_FROM_CHECKPOINT and os.path.exists(save_checkpoint_path):
model, optimizer, lr_scheduler, checkpoint_data = get_fasterrcnn_resnet50_fpn_v2_model(len(LABELS.keys()), save_checkpoint_path, DEVICE)
else:
model, optimizer, lr_scheduler, checkpoint_data = get_fasterrcnn_resnet50_fpn_v2_model(len(LABELS.keys()), None, DEVICE)
print(model)
Downloading: "https://download.pytorch.org/models/fasterrcnn_resnet50_fpn_v2_coco-dd69338a.pth" to /root/.cache/torch/hub/checkpoints/fasterrcnn_resnet50_fpn_v2_coco-dd69338a.pth 100%|██████████| 167M/167M [00:00<00:00, 194MB/s]
Cargando el modelo desde 0.
FasterRCNN(
(transform): GeneralizedRCNNTransform(
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
Resize(min_size=(800,), max_size=1333, mode='bilinear')
)
(backbone): BackboneWithFPN(
(body): IntermediateLayerGetter(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): Bottleneck(
(conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer2): Sequential(
(0): Bottleneck(
(conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer3): Sequential(
(0): Bottleneck(
(conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(5): Bottleneck(
(conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(layer4): Sequential(
(0): Bottleneck(
(conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): Bottleneck(
(conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
)
(fpn): FeaturePyramidNetwork(
(inner_blocks): ModuleList(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(2048, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(layer_blocks): ModuleList(
(0-3): 4 x Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(extra_blocks): LastLevelMaxPool()
)
)
(rpn): RegionProposalNetwork(
(anchor_generator): AnchorGenerator()
(head): RPNHead(
(conv): Sequential(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): ReLU(inplace=True)
)
)
(cls_logits): Conv2d(256, 3, kernel_size=(1, 1), stride=(1, 1))
(bbox_pred): Conv2d(256, 12, kernel_size=(1, 1), stride=(1, 1))
)
)
(roi_heads): RoIHeads(
(box_roi_pool): MultiScaleRoIAlign(featmap_names=['0', '1', '2', '3'], output_size=(7, 7), sampling_ratio=2)
(box_head): FastRCNNConvFCHead(
(0): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(1): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(2): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(3): Conv2dNormActivation(
(0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
)
(4): Flatten(start_dim=1, end_dim=-1)
(5): Linear(in_features=12544, out_features=1024, bias=True)
(6): ReLU(inplace=True)
)
(box_predictor): FastRCNNPredictor(
(cls_score): Linear(in_features=1024, out_features=2, bias=True)
(bbox_pred): Linear(in_features=1024, out_features=8, bias=True)
)
)
)
Podemos observar un resumen del mismo mediante torchinfo:
In [27]:
img, _ = test_dataset[0]
input_size = (1, *img.shape)
summary(model, input_size=input_size)
Out[27]:
==================================================================================================== Layer (type:depth-idx) Output Shape Param # ==================================================================================================== FasterRCNN [100, 4] -- ├─GeneralizedRCNNTransform: 1-1 [1, 3, 800, 1088] -- ├─BackboneWithFPN: 1-2 [1, 256, 13, 17] -- │ └─IntermediateLayerGetter: 2-1 [1, 2048, 25, 34] -- │ │ └─Conv2d: 3-1 [1, 64, 400, 544] (9,408) │ │ └─BatchNorm2d: 3-2 [1, 64, 400, 544] (128) │ │ └─ReLU: 3-3 [1, 64, 400, 544] -- │ │ └─MaxPool2d: 3-4 [1, 64, 200, 272] -- │ │ └─Sequential: 3-5 [1, 256, 200, 272] (215,808) │ │ └─Sequential: 3-6 [1, 512, 100, 136] 1,219,584 │ │ └─Sequential: 3-7 [1, 1024, 50, 68] 7,098,368 │ │ └─Sequential: 3-8 [1, 2048, 25, 34] 14,964,736 │ └─FeaturePyramidNetwork: 2-2 [1, 256, 13, 17] -- │ │ └─ModuleList: 3-15 -- (recursive) │ │ └─ModuleList: 3-16 -- (recursive) │ │ └─ModuleList: 3-15 -- (recursive) │ │ └─ModuleList: 3-16 -- (recursive) │ │ └─ModuleList: 3-15 -- (recursive) │ │ └─ModuleList: 3-16 -- (recursive) │ │ └─ModuleList: 3-15 -- (recursive) │ │ └─ModuleList: 3-16 -- (recursive) │ │ └─LastLevelMaxPool: 3-17 [1, 256, 200, 272] -- ├─RegionProposalNetwork: 1-3 [964, 4] -- │ └─RPNHead: 2-3 [1, 3, 200, 272] -- │ │ └─Sequential: 3-18 [1, 256, 200, 272] 1,180,160 │ │ └─Conv2d: 3-19 [1, 3, 200, 272] 771 │ │ └─Conv2d: 3-20 [1, 12, 200, 272] 3,084 │ │ └─Sequential: 3-21 [1, 256, 100, 136] (recursive) │ │ └─Conv2d: 3-22 [1, 3, 100, 136] (recursive) │ │ └─Conv2d: 3-23 [1, 12, 100, 136] (recursive) │ │ └─Sequential: 3-24 [1, 256, 50, 68] (recursive) │ │ └─Conv2d: 3-25 [1, 3, 50, 68] (recursive) │ │ └─Conv2d: 3-26 [1, 12, 50, 68] (recursive) │ │ └─Sequential: 3-27 [1, 256, 25, 34] (recursive) │ │ └─Conv2d: 3-28 [1, 3, 25, 34] (recursive) │ │ └─Conv2d: 3-29 [1, 12, 25, 34] (recursive) │ │ └─Sequential: 3-30 [1, 256, 13, 17] (recursive) │ │ └─Conv2d: 3-31 [1, 3, 13, 17] (recursive) │ │ └─Conv2d: 3-32 [1, 12, 13, 17] (recursive) │ └─AnchorGenerator: 2-4 [217413, 4] -- ├─RoIHeads: 1-4 [100, 4] -- │ └─MultiScaleRoIAlign: 2-5 [964, 256, 7, 7] -- │ └─FastRCNNConvFCHead: 2-6 [964, 1024] -- │ │ └─Conv2dNormActivation: 3-33 [964, 256, 7, 7] 590,336 │ │ └─Conv2dNormActivation: 3-34 [964, 256, 7, 7] 590,336 │ │ └─Conv2dNormActivation: 3-35 [964, 256, 7, 7] 590,336 │ │ └─Conv2dNormActivation: 3-36 [964, 256, 7, 7] 590,336 │ │ └─Flatten: 3-37 [964, 12544] -- │ │ └─Linear: 3-38 [964, 1024] 12,846,080 │ │ └─ReLU: 3-39 [964, 1024] -- │ └─FastRCNNPredictor: 2-7 [964, 2] -- │ │ └─Linear: 3-40 [964, 2] 2,050 │ │ └─Linear: 3-41 [964, 8] 8,200 ==================================================================================================== Total params: 43,256,153 Trainable params: 43,030,809 Non-trainable params: 225,344 Total mult-adds (G): 329.84 ==================================================================================================== Input size (MB): 1.44 Forward/backward pass size (MB): 4764.00 Params size (MB): 173.02 Estimated Total Size (MB): 4938.46 ====================================================================================================
In [28]:
# Limpiamos el chache.
clean_gpu_usage()
Initial GPU Usage | ID | GPU | MEM | ------------------ | 0 | 22% | 10% | GPU Usage after emptying the cache | ID | GPU | MEM | ------------------ | 0 | 10% | 3% |
Para comprobar el funcionamiento del modelo, podemos realizar un test de la inferencia y el entrenamiento:
In [29]:
data_loader = DataLoader(
train_dataset, batch_size=2, shuffle=True, collate_fn=lambda x: tuple(zip(*x))
)
model.to(DEVICE)
# Training
images, targets = next(iter(data_loader))
# Es necesario para que no compute los gradientes, sino queda "basura" en la GPU.
with torch.no_grad():
images = list(image.to(DEVICE) for image in images)
targets = [
{k: (v.to(DEVICE) if isinstance(v, torch.Tensor) else v) for k, v in t.items()}
for t in targets
]
output = model(images, targets) # Devuelve losses y detections
print(f"Training test: {output}")
# Inferencia
model.eval()
with torch.no_grad():
x = [torch.rand(3, 300, 400).to(DEVICE), torch.rand(3, 500, 400).to(DEVICE)]
predictions = model(x) # Returns predictions
print(f"Inference test: {predictions}")
Training test: {'loss_classifier': tensor(0.6804, device='cuda:0'), 'loss_box_reg': tensor(0.6662, device='cuda:0'), 'loss_objectness': tensor(2.6651, device='cuda:0'), 'loss_rpn_box_reg': tensor(0.1652, device='cuda:0')}
Inference test: [{'boxes': tensor([[1.1838e+00, 6.9196e+01, 2.3142e+01, 9.1906e+01],
[2.0914e+00, 1.9127e+00, 2.3696e+01, 2.3896e+01],
[1.1071e+00, 5.9000e+00, 2.9532e+01, 1.1002e+02],
[7.8116e-01, 3.8419e+01, 4.8221e+00, 2.1711e+02],
[2.0872e+00, 2.1879e+02, 1.7378e+01, 2.8514e+02],
[3.9836e-01, 1.3803e+01, 6.1088e+00, 1.0794e+02],
[4.2202e+00, 2.0602e+01, 7.8354e+01, 2.2338e+02],
[5.0875e-01, 9.7071e+01, 3.7838e+00, 3.0000e+02],
[3.3040e-01, 1.6243e+00, 2.4543e+00, 7.9984e+00],
[2.8347e+00, 2.3997e+02, 2.1734e+01, 2.9073e+02],
[1.0745e+01, 9.0602e+00, 1.0062e+02, 9.7987e+01],
[2.0819e+00, 2.2835e+01, 1.7662e+01, 9.3627e+01],
[1.3934e+00, 1.2516e+01, 4.2324e+01, 2.0761e+02],
[3.9529e+00, 1.9021e+02, 5.0760e+01, 2.4612e+02],
[2.7400e-01, 1.9149e+02, 1.3847e+01, 2.2384e+02],
[1.9286e+02, 1.3049e+02, 2.1644e+02, 1.5595e+02],
[8.5947e-02, 9.5661e+01, 1.2750e+00, 2.1012e+02],
[1.2327e+00, 9.4787e+00, 1.2566e+01, 2.4390e+02],
[3.7159e+00, 1.2878e+02, 3.7331e+01, 2.5260e+02],
[2.1700e+00, 1.6617e+02, 5.0782e+01, 2.2331e+02],
[3.5076e-02, 1.9495e+02, 1.2038e+00, 2.7788e+02],
[6.4926e+00, 3.0107e+00, 5.1604e+01, 2.3700e+01],
[4.0101e+00, 3.7491e+00, 5.4818e+01, 5.5457e+01],
[1.4809e+00, 1.0776e+01, 1.7659e+01, 4.7967e+01],
[7.2419e-02, 1.4596e+02, 1.0681e+00, 2.6651e+02],
[2.6972e-01, 3.9373e+00, 2.5625e+00, 1.2599e+01],
[4.6435e-01, 1.3784e+01, 1.6058e+00, 1.1518e+02],
[2.9308e-01, 6.4677e+01, 1.7640e+00, 1.7120e+02],
[3.5298e+00, 9.0067e+01, 3.5453e+01, 2.0475e+02],
[5.6804e-02, 2.7477e+01, 1.0741e+00, 1.2423e+02],
[3.2147e+02, 1.7071e+02, 3.4544e+02, 1.9476e+02],
[3.6411e+00, 7.3655e+00, 3.6200e+01, 7.0725e+01],
[2.5458e+00, 1.9716e+02, 1.7516e+01, 2.5536e+02],
[5.4481e+00, 6.9872e+01, 5.1643e+01, 1.2042e+02],
[1.0716e+02, 1.0917e+02, 1.1169e+02, 1.1361e+02],
[2.1038e+00, 1.9312e+01, 1.5508e+01, 1.6179e+02],
[2.9056e+00, 3.5271e+01, 3.2552e+01, 1.5827e+02],
[6.4408e+00, 1.6875e+00, 9.1419e+01, 1.9765e+01],
[1.5883e+00, 2.7284e+02, 2.5581e+01, 2.9790e+02],
[2.9982e+00, 9.5912e+01, 1.5944e+01, 1.6825e+02],
[6.3398e+00, 1.0082e+00, 6.1967e+01, 9.3382e+00],
[3.8002e+00, 5.1851e-01, 1.0214e+02, 1.0432e+01],
[4.3802e-01, 1.8655e+00, 3.5049e+00, 1.1225e+01],
[1.0822e+02, 1.1025e+02, 1.1236e+02, 1.1483e+02],
[1.7376e+02, 1.1541e+02, 2.0003e+02, 1.4017e+02],
[1.0701e+01, 6.7086e+00, 1.3412e+02, 1.3945e+02],
[1.4562e+00, 4.4613e-01, 9.1344e+00, 3.0569e+00],
[3.8447e+00, 1.6696e+02, 1.5122e+01, 2.4004e+02],
[9.9593e+00, 1.3868e+02, 9.0940e+01, 2.3077e+02],
[2.7560e-02, 5.4459e+00, 1.1902e+00, 8.7901e+01],
[9.4716e+01, 7.0194e-01, 1.6560e+02, 1.6746e+00],
[3.8166e+00, 2.2367e+02, 3.7211e+01, 2.9039e+02],
[1.3099e-02, 4.9762e+01, 1.1817e+00, 1.7369e+02],
[4.1312e+00, 6.8820e+01, 1.5107e+01, 1.4141e+02],
[3.1483e+02, 1.7319e+02, 3.3661e+02, 1.9594e+02],
[6.2488e-01, 7.1906e-01, 6.6207e+00, 3.5269e+00],
[3.7218e+00, 1.4291e+02, 5.3254e+01, 1.9882e+02],
[3.3209e+00, 1.3486e+02, 1.5242e+01, 2.1622e+02],
[1.0595e+01, 4.6290e+00, 1.0690e+02, 3.7243e+01],
[5.3710e+00, 1.6236e+00, 3.8800e+01, 1.8306e+01],
[3.4794e+00, 5.5858e+01, 2.0158e+01, 1.1845e+02],
[3.1076e-01, 4.3464e+00, 1.8500e+00, 6.1777e+01],
[1.6865e+02, 1.2305e+02, 1.7173e+02, 1.2590e+02],
[1.7586e-01, 1.8609e+02, 4.8988e+00, 3.0000e+02],
[1.4805e+00, 1.3130e+02, 1.3113e+01, 2.8720e+02],
[3.3126e+02, 1.1172e+02, 3.3665e+02, 1.1522e+02],
[1.1555e-01, 6.8069e+00, 2.1260e+00, 2.4083e+01],
[3.6034e+02, 1.8933e+00, 4.0000e+02, 3.4411e+01],
[1.3855e+01, 7.7996e+01, 1.8602e+01, 8.1664e+01],
[2.1527e+02, 1.0915e-01, 2.9274e+02, 1.2658e+00],
[1.6683e+00, 4.0281e-01, 8.5713e+01, 5.6490e+00],
[3.1014e+02, 2.1185e+00, 4.0000e+02, 1.8870e+01],
[3.9484e+01, 1.0288e+00, 2.1749e+02, 4.2918e+00],
[1.7647e+02, 1.2735e+02, 1.8039e+02, 1.3060e+02],
[1.6185e+02, 1.7766e+02, 1.6584e+02, 1.8220e+02],
[1.4054e+01, 8.5308e+01, 1.5412e+02, 2.1644e+02],
[2.0078e-01, 2.3431e+02, 1.7856e+00, 3.0000e+02],
[3.3376e+02, 9.4217e+01, 3.5795e+02, 1.1706e+02],
[3.4937e+02, 1.5238e-01, 4.0000e+02, 1.8748e+00],
[2.5362e+02, 1.8775e-01, 3.2085e+02, 1.2539e+00],
[1.4355e+02, 2.2069e+02, 1.4756e+02, 2.2423e+02],
[1.9599e+02, 1.3411e+02, 2.2511e+02, 1.6347e+02],
[5.7509e-01, 2.4297e-01, 3.1842e+00, 4.9793e+00],
[1.8273e+02, 1.3258e+02, 2.1365e+02, 1.6364e+02],
[1.5874e+01, 7.4106e+01, 2.0625e+01, 7.7199e+01],
[1.7554e+02, 1.2678e+02, 1.7852e+02, 1.3089e+02],
[2.8416e+02, 1.9800e+02, 3.2428e+02, 2.1737e+02],
[2.9554e+00, 2.7059e-01, 5.3310e+01, 2.2339e+00],
[3.2463e+02, 1.2157e+00, 4.0000e+02, 5.4937e+00],
[3.0187e+00, 2.3788e-01, 1.1854e+01, 2.3808e+00],
[3.8098e+02, 9.1978e+00, 4.0000e+02, 5.8569e+01],
[1.9829e+02, 1.5134e+01, 2.9126e+02, 1.0151e+02],
[1.6367e+02, 1.1864e+02, 1.9315e+02, 1.4537e+02],
[1.7011e+02, 1.2165e+02, 1.7372e+02, 1.2477e+02],
[1.7569e+02, 1.2421e+02, 1.7872e+02, 1.2812e+02],
[1.3117e+01, 7.6318e+01, 1.8046e+01, 7.9700e+01],
[5.4787e-02, 3.8313e-01, 1.5586e+00, 6.9443e+00],
[6.1164e-02, 2.2691e+00, 1.6730e+00, 1.0278e+01],
[1.7136e+02, 1.3245e+02, 2.4726e+02, 1.6797e+02],
[1.5230e-01, 4.4514e+00, 3.0023e+00, 2.0281e+01]], device='cuda:0'), 'labels': tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
device='cuda:0'), 'scores': tensor([0.6048, 0.5852, 0.5780, 0.5773, 0.5673, 0.5620, 0.5609, 0.5598, 0.5587, 0.5543, 0.5524, 0.5509, 0.5493, 0.5489, 0.5469, 0.5449, 0.5442, 0.5437, 0.5416, 0.5406, 0.5401, 0.5397, 0.5394, 0.5392, 0.5388, 0.5374, 0.5374, 0.5361, 0.5357, 0.5338, 0.5332, 0.5319, 0.5318, 0.5315, 0.5304, 0.5276, 0.5267, 0.5261, 0.5255,
0.5249, 0.5248, 0.5244, 0.5229, 0.5228, 0.5221, 0.5219, 0.5210, 0.5206, 0.5204, 0.5201, 0.5193, 0.5189, 0.5186, 0.5184, 0.5174, 0.5173, 0.5165, 0.5161, 0.5153, 0.5145, 0.5139, 0.5115, 0.5112, 0.5111, 0.5106, 0.5103, 0.5085, 0.5084, 0.5082, 0.5081, 0.5080, 0.5076, 0.5065, 0.5055, 0.5047, 0.5047, 0.5037, 0.5031,
0.5024, 0.5012, 0.5010, 0.5007, 0.5000, 0.4995, 0.4993, 0.4988, 0.4982, 0.4979, 0.4965, 0.4957, 0.4946, 0.4935, 0.4935, 0.4933, 0.4931, 0.4923, 0.4918, 0.4918, 0.4916, 0.4916], device='cuda:0')}, {'boxes': tensor([[2.9572e+00, 2.4565e+00, 3.2190e+01, 3.1069e+01],
[2.0088e+02, 2.5354e+02, 2.0661e+02, 2.5716e+02],
[3.0846e+00, 1.4146e+02, 3.2825e+01, 1.7229e+02],
[2.1422e+00, 6.9173e+00, 4.1508e+01, 1.3565e+02],
[5.9225e+00, 3.0755e+00, 6.5494e+01, 5.7274e+01],
[4.8096e+00, 1.1340e+01, 7.2981e+01, 9.7655e+01],
[1.5955e+01, 1.5767e+01, 1.3224e+02, 1.3304e+02],
[6.7925e-01, 9.0284e+01, 5.0381e+00, 3.6284e+02],
[8.4425e-01, 2.6454e+02, 1.7244e+01, 3.9061e+02],
[6.6205e-01, 1.0590e+01, 6.7564e+00, 2.1159e+02],
[2.1201e+01, 5.2711e+00, 2.2609e+02, 1.4988e+02],
[6.2056e-01, 1.7410e+02, 4.7024e+00, 4.6401e+02],
[7.9431e-01, 2.0826e+01, 8.7202e+00, 1.3981e+02],
[2.0536e+00, 3.2215e-01, 7.1818e+01, 7.3487e+00],
[6.7951e-01, 1.9672e+02, 2.1081e+01, 3.2557e+02],
[1.5797e+00, 1.0635e+01, 1.5725e+01, 3.4195e+02],
[3.2884e+00, 4.2028e+02, 2.4966e+01, 4.8669e+02],
[1.9881e+02, 2.7372e+02, 2.0474e+02, 2.7812e+02],
[9.4354e+00, 2.1675e+00, 8.3125e+01, 1.3218e+01],
[1.4782e+02, 3.0753e+02, 1.5338e+02, 3.1153e+02],
[2.0178e+02, 2.5219e+02, 2.0625e+02, 2.5536e+02],
[6.7084e-01, 1.0757e+02, 2.0900e+00, 2.6625e+02],
[2.7756e+00, 1.6120e+01, 3.0579e+01, 8.8541e+01],
[3.2782e+00, 2.7495e+01, 2.2096e+01, 1.9295e+02],
[4.8128e-02, 7.9474e+01, 1.5266e+00, 2.3062e+02],
[4.5666e-01, 3.0428e+02, 1.7345e+00, 4.6057e+02],
[5.7889e-02, 9.5392e+00, 5.0852e+00, 1.4950e+02],
[5.5091e-01, 2.9246e+01, 1.8917e+00, 1.8952e+02],
[1.4228e+00, 3.8099e+02, 8.1706e+00, 4.9336e+02],
[1.2263e-01, 3.2505e+02, 1.3434e+00, 4.8270e+02],
[2.7565e+01, 1.3733e+01, 1.7493e+02, 2.3225e+02],
[7.0036e-01, 3.8301e+02, 2.4344e+00, 5.0000e+02],
[1.7695e-01, 2.1097e+02, 1.4366e+00, 3.6848e+02],
[6.9054e-02, 1.2689e+02, 1.4101e+00, 2.7683e+02],
[8.1419e+00, 3.2950e+01, 1.0311e+02, 2.9950e+02],
[4.8131e+00, 1.2541e+02, 2.2371e+01, 2.1387e+02],
[4.5420e-01, 2.2778e+02, 1.8244e+00, 3.9729e+02],
[3.6512e-01, 1.5571e+01, 2.0793e+00, 1.2108e+02],
[1.1027e+01, 4.9755e+01, 1.2915e+02, 1.9833e+02],
[1.4511e-01, 1.6314e+02, 1.5909e+00, 3.2323e+02],
[1.9906e+02, 2.7244e+02, 2.0317e+02, 2.7682e+02],
[6.5839e+00, 5.4642e+01, 6.7739e+01, 1.2650e+02],
[1.9987e+02, 2.5426e+02, 2.0485e+02, 2.5779e+02],
[2.9493e+02, 8.1122e+01, 3.4949e+02, 1.0579e+02],
[5.5140e+00, 2.9824e+02, 1.9822e+01, 4.0786e+02],
[3.7835e+00, 5.0734e+01, 4.8852e+01, 1.9669e+02],
[9.4413e+00, 7.1423e-01, 2.0790e+02, 1.5162e+01],
[1.6277e+00, 8.9502e+01, 3.4839e+01, 2.1772e+02],
[5.1889e-02, 1.5333e+01, 1.4428e+00, 1.3498e+02],
[2.0047e+02, 2.5229e+02, 2.0483e+02, 2.5551e+02],
[2.3069e+02, 2.4127e+02, 2.3474e+02, 2.4578e+02],
[1.6891e+02, 2.5262e+02, 2.0098e+02, 2.8441e+02],
[1.4861e+02, 3.0908e+02, 1.5482e+02, 3.1305e+02],
[2.3168e+02, 2.4066e+02, 2.3591e+02, 2.4487e+02],
[1.4790e+02, 1.9031e+02, 2.0305e+02, 2.4863e+02],
[1.1457e+02, 5.0807e+00, 2.2678e+02, 8.8072e+01],
[7.7154e-02, 3.9603e+02, 1.5485e+00, 5.0000e+02],
[4.5597e-02, 2.5482e+02, 1.0904e+00, 4.5117e+02],
[2.3167e+02, 2.3934e+02, 2.3624e+02, 2.4327e+02],
[7.2726e+00, 1.3105e+02, 6.7021e+01, 1.9353e+02],
[1.9821e+02, 2.7403e+02, 2.0303e+02, 2.7945e+02],
[1.7079e-01, 3.9485e+02, 5.0123e+00, 5.0000e+02],
[8.5375e-01, 3.0523e+02, 6.2928e+00, 5.0000e+02],
[1.0384e+00, 4.9475e+01, 2.6772e+01, 3.8799e+02],
[2.2727e-02, 1.9030e+02, 1.0879e+00, 3.4134e+02],
[8.5759e-01, 4.0329e+02, 1.6060e+01, 5.0000e+02],
[1.9978e+02, 2.5134e+02, 2.0360e+02, 2.5513e+02],
[5.4344e+00, 2.4743e+02, 2.1041e+01, 3.4845e+02],
[8.2601e+00, 2.4818e+02, 7.1193e+01, 3.1724e+02],
[1.0422e+01, 4.8586e+00, 7.2868e+01, 2.9846e+01],
[1.2046e+01, 3.0324e+00, 1.2684e+02, 2.8320e+01],
[1.9960e+02, 2.5301e+02, 2.0417e+02, 2.5659e+02],
[3.2825e+00, 6.5435e+01, 6.2687e+01, 3.6329e+02],
[5.1014e+00, 1.7162e+02, 2.1501e+01, 2.6657e+02],
[2.3489e+01, 2.8921e+01, 7.3234e+01, 1.4216e+02],
[1.0540e+02, 3.2902e+02, 1.6594e+02, 3.9463e+02],
[6.5591e+00, 1.8629e+02, 4.7375e+01, 3.2890e+02],
[1.4979e+02, 3.0992e+02, 1.5474e+02, 3.1525e+02],
[8.7145e+00, 1.9857e+02, 7.2090e+01, 2.6936e+02],
[4.3641e+00, 9.7006e-01, 1.6513e+02, 7.6226e+00],
[1.9751e+02, 2.7181e+02, 2.0187e+02, 2.7538e+02],
[1.9659e+02, 2.8176e+02, 2.0138e+02, 2.8552e+02],
[1.9455e+02, 2.7586e+02, 1.9912e+02, 2.8029e+02],
[4.7114e+00, 3.4956e+02, 2.2535e+01, 4.7276e+02],
[1.3162e+01, 1.9603e+02, 1.2805e+02, 3.2259e+02],
[1.5252e+02, 2.1008e+02, 1.8703e+02, 2.4498e+02],
[1.1041e+02, 4.5710e-01, 3.1646e+02, 5.4713e+00],
[1.5395e+02, 4.9900e-01, 4.0000e+02, 1.1191e+01],
[1.2606e+00, 4.6167e+02, 2.4622e+01, 5.0000e+02],
[3.1327e+02, 7.3127e-01, 4.0000e+02, 6.9836e+00],
[1.2449e+02, 2.9084e+02, 1.4975e+02, 3.4894e+02],
[2.0316e+02, 2.5234e+02, 2.0736e+02, 2.5603e+02],
[6.2881e+00, 1.0501e+02, 4.6764e+01, 2.4712e+02],
[1.9958e+02, 2.8225e+02, 2.0442e+02, 2.8616e+02],
[1.9600e+02, 2.7538e+02, 2.0003e+02, 2.8076e+02],
[7.0758e+00, 9.1582e+01, 6.5400e+01, 1.5080e+02],
[8.9005e+01, 7.9374e+00, 3.9246e+02, 1.2294e+02],
[3.3688e+02, 5.7885e+00, 4.0000e+02, 6.9993e+01],
[3.7779e+02, 2.6538e+01, 3.9873e+02, 9.5263e+01],
[2.6669e+02, 7.5127e+00, 3.9299e+02, 4.4415e+01]], device='cuda:0'), 'labels': tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
device='cuda:0'), 'scores': tensor([0.5589, 0.5493, 0.5484, 0.5391, 0.5372, 0.5351, 0.5336, 0.5296, 0.5286, 0.5280, 0.5243, 0.5218, 0.5189, 0.5188, 0.5187, 0.5178, 0.5151, 0.5146, 0.5140, 0.5139, 0.5138, 0.5137, 0.5133, 0.5129, 0.5116, 0.5110, 0.5096, 0.5095, 0.5090, 0.5077, 0.5076, 0.5074, 0.5071, 0.5069, 0.5060, 0.5049, 0.5035, 0.5032, 0.5031,
0.5030, 0.5027, 0.5026, 0.5025, 0.5012, 0.5010, 0.5006, 0.5000, 0.4998, 0.4997, 0.4976, 0.4975, 0.4971, 0.4963, 0.4958, 0.4957, 0.4948, 0.4945, 0.4942, 0.4942, 0.4932, 0.4930, 0.4929, 0.4923, 0.4920, 0.4916, 0.4912, 0.4907, 0.4906, 0.4899, 0.4898, 0.4890, 0.4890, 0.4888, 0.4874, 0.4873, 0.4870, 0.4868, 0.4863,
0.4861, 0.4842, 0.4829, 0.4817, 0.4814, 0.4810, 0.4806, 0.4805, 0.4804, 0.4794, 0.4788, 0.4783, 0.4779, 0.4776, 0.4773, 0.4769, 0.4765, 0.4761, 0.4761, 0.4761, 0.4756, 0.4754], device='cuda:0')}]
In [30]:
# Limpiamos el chache.
clean_gpu_usage()
Initial GPU Usage | ID | GPU | MEM | ------------------ | 0 | 93% | 25% | GPU Usage after emptying the cache | ID | GPU | MEM | ------------------ | 0 | 36% | 3% |
Entrenar el modelo¶
Clase que se utiliza para detener el entrenamiento si no mejora la métrica (pérdida en evaluación) durante N épocas.
In [31]:
class EarlyStopper:
def __init__(self, patience: int = 1, min_delta: float = 0):
self.patience = patience
self.min_delta = min_delta
self.counter = 0
self.min_validation_loss = float("inf")
def early_stop(self, validation_loss: float) -> bool:
if validation_loss < self.min_validation_loss:
self.min_validation_loss = validation_loss
self.counter = 0
elif validation_loss >= (self.min_validation_loss + self.min_delta):
self.counter += 1
if self.counter >= self.patience:
return True
return False
Wrapper sobre tran_one_epoch (permite realizar el entrenamiento en una época) de las utilidades de entrenamiento de pytorch/resources/vision/detection:
In [32]:
def coco_train(
model: nn.Module,
data_loader: DataLoader,
device: str,
optimizer: optim.Optimizer,
epoch: int = None,
criterion: nn.Module = None,
verbose: bool = True,
) -> tuple[float, float]:
model.to(device)
train_loss = 0.0
train_acc = 0.0
metric_logger: MetricLogger = train_one_epoch(
model, optimizer, data_loader, DEVICE, epoch, print_freq=10
)
train_loss = metric_logger.meters["loss"].global_avg # Obtener la pérdida promedio
return train_loss, train_acc
Wrapper sobre evaluate (permite evaluar utilizando el formato de COCO -se dividen entre los tamaños de los objetos-) de las utilidades de entrenamiento de pytorch/resources/vision/detection:
In [33]:
def coco_evaluate(
model: nn.Module,
data_loader: DataLoader,
device: str,
optimizer: optim.Optimizer = None,
epoch: int = None,
criterion: nn.Module = None,
verbose: bool = True,
) -> tuple[float, float]:
model.to(device)
test_loss = 0.0
test_acc = 0.0
coco_evaluator: CocoEvaluator = evaluate(model, data_loader, device)
test_loss = coco_evaluator.coco_eval["bbox"].stats[0]
test_acc = coco_evaluator.coco_eval["bbox"].stats[1]
return test_loss, test_acc
Función que permite realizar el entrenamiento utilizando las funcionalidades de COCO. Utiliza EarlyStopping y también continúa el entrenamiento desde el útlimo checkpoint:
In [34]:
def fit(
model: nn.Module,
train_dataloader: DataLoader,
test_dataloader: DataLoader,
optimizer: optim.Optimizer,
criterion: nn.Module,
scheduler: optim.Optimizer = None,
early_stopping: EarlyStopper = None,
epochs: int = EPOCHS,
verbose: bool = True,
device: str = DEVICE,
start_from_checkpoint_data: Dict = None,
base_dir: str = None,
model_id: str = None,
save_checkpoint: bool = False,
save_best_model: bool = False,
save_results: bool = False
) -> Tuple[List[float], List[float], List[float], List[float]]:
# Chequeos iniciales
if (save_best_model or save_checkpoint or save_results) and not (base_dir and model_id):
raise ValueError("base_dir y model_id no pueden ser vacíos si save_best_model o save_checkpoint son verdaderos.")
if start_from_checkpoint_data:
starting_epoch = start_from_checkpoint_data['epoch'] + 1
if starting_epoch > epochs:
raise IndexError(f"Mal inicio desde checkpoint: starting_epoch > epochs")
print(f"Traning from checkpoint at epoch {starting_epoch}\n")
else:
starting_epoch = 1
print("Training... \n")
train_losses = [] if not start_from_checkpoint_data else start_from_checkpoint_data['train_losses']
test_losses = [] if not start_from_checkpoint_data else start_from_checkpoint_data['test_losses']
train_accs = [] if not start_from_checkpoint_data else start_from_checkpoint_data['train_accs']
test_accs = [] if not start_from_checkpoint_data else start_from_checkpoint_data['test_accs']
best_test_acc = -float('inf') if not start_from_checkpoint_data else start_from_checkpoint_data['"best_test_acc"']
best_test_acc = -float('inf')
pbar = trange(starting_epoch, epochs + 1, desc="Epoch", unit="epoch", colour="green")
for epoch in pbar:
train_loss, train_acc = coco_train(
model=model,
data_loader=train_dataloader,
optimizer=optimizer,
device=device,
epoch=epoch,
criterion=None,
verbose=verbose
)
test_loss, test_acc = coco_evaluate(model=model,
data_loader=test_dataloader,
device=device,
verbose=verbose)
if scheduler: scheduler.step()
train_losses.append(train_loss)
test_losses.append(test_loss)
train_accs.append(train_acc)
test_accs.append(test_acc)
pbar.set_postfix({
"Train Loss": format(train_loss, ".3f"),
"Test Loss": format(test_loss, ".3f"),
"Train Accuracy": format(train_acc, ".3f"),
"Test Accuracy": format(test_acc, ".3f"),
})
# Checkpoint
if save_checkpoint:
file_name_path = os.path.join(base_dir, model_id + '_checkpoint.tar')
os.makedirs(base_dir, exist_ok=True)
torch.save(
{
"model_state_dict": model.state_dict(),
"optimizer_state_dict": optimizer.state_dict(),
"scheduler_state_dict": scheduler.state_dict(),
"data": {
"epoch": epoch,
"train_losses": train_losses,
"test_losses": test_losses,
"train_accs": train_accs,
"test_accs": test_accs,
"best_test_acc": best_test_acc
},
},
file_name_path,
)
# Save best model
if save_best_model and test_acc > best_test_acc:
best_test_acc = test_acc
file_name_path = os.path.join(base_dir, model_id + '_best_model.pt')
os.makedirs(base_dir, exist_ok=True)
torch.save(model.state_dict(), file_name_path)
if verbose:
print(f"Best model saved at epoch {epoch}!")
# Early stopping
if early_stopping and early_stopping.early_stop(test_loss):
print("\nEarly stopping activated.\n")
break
# Save results in JSON
if save_results:
results = {
"train_losses": train_losses,
"test_losses": test_losses,
"train_accs": train_accs,
"test_accs": test_accs,
"best_test_acc": best_test_acc
}
results_file_path = os.path.join(base_dir, model_id + '_results.json')
with open(results_file_path, 'w') as json_file:
json.dump(results, json_file, indent=4)
if verbose:
print(f"Results saved to {results_file_path}!")
print("\n \nFinished Training")
return train_losses, test_losses, train_accs, test_accs
Una vez que tenemos todo configurado, realizamos el entrenamiento:
In [35]:
EPOCHS = 20
BATCH_SIZE = 4
train_dataloader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True, collate_fn=lambda x: tuple(zip(*x)))
test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False, collate_fn=lambda x: tuple(zip(*x)))
train_losses, test_losses, train_accs, test_accs = fit(
model=model,
train_dataloader=train_dataloader,
test_dataloader=test_dataloader,
optimizer=optimizer,
criterion=None,
scheduler=lr_scheduler,
early_stopping=EarlyStopper(5, 0),
epochs=EPOCHS,
verbose=False,
device=DEVICE,
start_from_checkpoint_data=checkpoint_data,
base_dir=os.path.join(
RUTA_CARPETA_MODELS,
FASTERRCNN_RESNET50_FPN_V2_MODEL_ID),
model_id=FASTERRCNN_RESNET50_FPN_V2_MODEL_ID,
save_checkpoint=LOAD_MODEL_FROM_CHECKPOINT,
save_best_model = True,
save_results = True
)
Training...
Epoch: 0%| | 0/20 [00:00<?, ?epoch/s]/kaggle/working/resources/utils/engine.py:30: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with torch.cuda.amp.autocast(enabled=scaler is not None):
Epoch: [1] [ 0/70] eta: 0:17:38 lr: 0.005000 loss: 3.2332 (3.2332) loss_classifier: 0.6616 (0.6616) loss_box_reg: 0.5221 (0.5221) loss_objectness: 1.9527 (1.9527) loss_rpn_box_reg: 0.0968 (0.0968) time: 15.1189 data: 14.0877 max mem: 7857 Epoch: [1] [10/70] eta: 0:15:42 lr: 0.005000 loss: 1.4919 (1.7196) loss_classifier: 0.4717 (0.4635) loss_box_reg: 0.5568 (0.5668) loss_objectness: 0.4557 (0.6076) loss_rpn_box_reg: 0.0749 (0.0816) time: 15.7142 data: 14.9562 max mem: 8029 Epoch: [1] [20/70] eta: 0:13:08 lr: 0.005000 loss: 1.2126 (1.3953) loss_classifier: 0.3121 (0.3739) loss_box_reg: 0.5044 (0.5462) loss_objectness: 0.2501 (0.4009) loss_rpn_box_reg: 0.0551 (0.0743) time: 15.8020 data: 15.0573 max mem: 8029 Epoch: [1] [30/70] eta: 0:10:16 lr: 0.005000 loss: 0.8895 (1.1661) loss_classifier: 0.2166 (0.3068) loss_box_reg: 0.4733 (0.4921) loss_objectness: 0.1004 (0.3014) loss_rpn_box_reg: 0.0462 (0.0657) time: 15.2296 data: 14.4679 max mem: 8029 Epoch: [1] [40/70] eta: 0:07:50 lr: 0.005000 loss: 0.6386 (1.0332) loss_classifier: 0.1545 (0.2675) loss_box_reg: 0.3661 (0.4596) loss_objectness: 0.0543 (0.2459) loss_rpn_box_reg: 0.0424 (0.0601) time: 15.5631 data: 14.8146 max mem: 8029 Epoch: [1] [50/70] eta: 0:05:14 lr: 0.005000 loss: 0.5859 (0.9429) loss_classifier: 0.1366 (0.2417) loss_box_reg: 0.3462 (0.4324) loss_objectness: 0.0698 (0.2116) loss_rpn_box_reg: 0.0456 (0.0572) time: 16.1752 data: 15.4312 max mem: 8029 Epoch: [1] [60/70] eta: 0:02:34 lr: 0.005000 loss: 0.5046 (0.8723) loss_classifier: 0.1232 (0.2251) loss_box_reg: 0.2963 (0.4094) loss_objectness: 0.0452 (0.1844) loss_rpn_box_reg: 0.0343 (0.0534) time: 15.0471 data: 14.2814 max mem: 8029 Epoch: [1] [69/70] eta: 0:00:15 lr: 0.005000 loss: 0.5259 (0.8405) loss_classifier: 0.1410 (0.2160) loss_box_reg: 0.3144 (0.4018) loss_objectness: 0.0351 (0.1689) loss_rpn_box_reg: 0.0389 (0.0538) time: 15.1424 data: 14.3768 max mem: 8029 Epoch: [1] Total time: 0:18:05 (15.5005 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3895 (0.3895) evaluator_time: 0.9710 (0.9710) time: 2.8201 data: 1.4578 max mem: 8029
Epoch: 0%| | 0/20 [18:56<?, ?epoch/s, Train Loss=0.841, Test Loss=0.519, Train Accuracy=0.000, Test Accuracy=0.939]
Test: [17/18] eta: 0:00:01 model_time: 0.3760 (0.3726) evaluator_time: 0.2231 (0.2840) time: 1.7438 data: 1.0852 max mem: 8029 Test: Total time: 0:00:31 (1.7439 s / it) Averaged stats: model_time: 0.3760 (0.3726) evaluator_time: 0.2231 (0.2840) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.519 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.939 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.525 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.603 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.018 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.174 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.602 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.057 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.604
Epoch: 5%|▌ | 1/20 [18:57<6:00:09, 1137.34s/epoch, Train Loss=0.841, Test Loss=0.519, Train Accuracy=0.000, Test Accuracy=0.939]
Epoch: [2] [ 0/70] eta: 0:14:26 lr: 0.005000 loss: 0.3349 (0.3349) loss_classifier: 0.1098 (0.1098) loss_box_reg: 0.1718 (0.1718) loss_objectness: 0.0387 (0.0387) loss_rpn_box_reg: 0.0146 (0.0146) time: 12.3805 data: 11.5911 max mem: 8029 Epoch: [2] [10/70] eta: 0:16:21 lr: 0.005000 loss: 0.5340 (0.5607) loss_classifier: 0.1341 (0.1472) loss_box_reg: 0.3160 (0.3241) loss_objectness: 0.0387 (0.0441) loss_rpn_box_reg: 0.0451 (0.0454) time: 16.3650 data: 15.6016 max mem: 8029 Epoch: [2] [20/70] eta: 0:12:46 lr: 0.005000 loss: 0.5230 (0.5194) loss_classifier: 0.1341 (0.1370) loss_box_reg: 0.3067 (0.2958) loss_objectness: 0.0385 (0.0447) loss_rpn_box_reg: 0.0392 (0.0419) time: 15.4750 data: 14.7076 max mem: 8029 Epoch: [2] [30/70] eta: 0:10:21 lr: 0.005000 loss: 0.4946 (0.5198) loss_classifier: 0.1371 (0.1345) loss_box_reg: 0.2767 (0.3028) loss_objectness: 0.0366 (0.0417) loss_rpn_box_reg: 0.0342 (0.0408) time: 15.0846 data: 14.3130 max mem: 8029 Epoch: [2] [40/70] eta: 0:07:43 lr: 0.005000 loss: 0.5534 (0.5172) loss_classifier: 0.1282 (0.1303) loss_box_reg: 0.3298 (0.3046) loss_objectness: 0.0366 (0.0409) loss_rpn_box_reg: 0.0384 (0.0414) time: 15.5668 data: 14.8054 max mem: 8029 Epoch: [2] [50/70] eta: 0:05:07 lr: 0.005000 loss: 0.4957 (0.5097) loss_classifier: 0.1157 (0.1278) loss_box_reg: 0.3025 (0.3022) loss_objectness: 0.0258 (0.0393) loss_rpn_box_reg: 0.0371 (0.0405) time: 15.1364 data: 14.3765 max mem: 8029 Epoch: [2] [60/70] eta: 0:02:34 lr: 0.005000 loss: 0.4740 (0.5043) loss_classifier: 0.1135 (0.1276) loss_box_reg: 0.2922 (0.2986) loss_objectness: 0.0258 (0.0384) loss_rpn_box_reg: 0.0333 (0.0397) time: 15.4349 data: 14.6800 max mem: 8030 Epoch: [2] [69/70] eta: 0:00:15 lr: 0.005000 loss: 0.4287 (0.4996) loss_classifier: 0.1134 (0.1266) loss_box_reg: 0.2804 (0.2973) loss_objectness: 0.0233 (0.0365) loss_rpn_box_reg: 0.0319 (0.0393) time: 14.8687 data: 14.1221 max mem: 8030 Epoch: [2] Total time: 0:17:51 (15.3026 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:52 model_time: 0.3858 (0.3858) evaluator_time: 0.9602 (0.9602) time: 2.8910 data: 1.5433 max mem: 8030
Epoch: 5%|▌ | 1/20 [37:41<6:00:09, 1137.34s/epoch, Train Loss=0.500, Test Loss=0.573, Train Accuracy=0.000, Test Accuracy=0.948]
Test: [17/18] eta: 0:00:01 model_time: 0.3776 (0.3743) evaluator_time: 0.2167 (0.2716) time: 1.7857 data: 1.1378 max mem: 8030 Test: Total time: 0:00:32 (1.7858 s / it) Averaged stats: model_time: 0.3776 (0.3743) evaluator_time: 0.2167 (0.2716) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.573 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.948 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.643 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.001 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.644 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.190 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.641 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.057 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.643
Epoch: 10%|█ | 2/20 [37:42<5:39:01, 1130.11s/epoch, Train Loss=0.500, Test Loss=0.573, Train Accuracy=0.000, Test Accuracy=0.948]
Epoch: [3] [ 0/70] eta: 0:15:51 lr: 0.005000 loss: 0.4577 (0.4577) loss_classifier: 0.1326 (0.1326) loss_box_reg: 0.2820 (0.2820) loss_objectness: 0.0178 (0.0178) loss_rpn_box_reg: 0.0252 (0.0252) time: 13.5945 data: 12.7850 max mem: 8030 Epoch: [3] [10/70] eta: 0:14:57 lr: 0.005000 loss: 0.4520 (0.4416) loss_classifier: 0.1279 (0.1168) loss_box_reg: 0.2820 (0.2662) loss_objectness: 0.0209 (0.0254) loss_rpn_box_reg: 0.0286 (0.0332) time: 14.9520 data: 14.1719 max mem: 8030 Epoch: [3] [20/70] eta: 0:12:46 lr: 0.005000 loss: 0.4364 (0.4374) loss_classifier: 0.1161 (0.1133) loss_box_reg: 0.2754 (0.2663) loss_objectness: 0.0209 (0.0237) loss_rpn_box_reg: 0.0318 (0.0342) time: 15.4133 data: 14.6347 max mem: 8030 Epoch: [3] [30/70] eta: 0:10:18 lr: 0.005000 loss: 0.4135 (0.4133) loss_classifier: 0.1025 (0.1075) loss_box_reg: 0.2627 (0.2530) loss_objectness: 0.0188 (0.0209) loss_rpn_box_reg: 0.0290 (0.0319) time: 15.7276 data: 14.9473 max mem: 8030 Epoch: [3] [40/70] eta: 0:07:49 lr: 0.005000 loss: 0.3894 (0.4185) loss_classifier: 0.1029 (0.1074) loss_box_reg: 0.2414 (0.2583) loss_objectness: 0.0160 (0.0212) loss_rpn_box_reg: 0.0275 (0.0316) time: 15.9733 data: 15.2049 max mem: 8030 Epoch: [3] [50/70] eta: 0:05:14 lr: 0.005000 loss: 0.4335 (0.4227) loss_classifier: 0.1076 (0.1092) loss_box_reg: 0.2634 (0.2593) loss_objectness: 0.0185 (0.0222) loss_rpn_box_reg: 0.0256 (0.0319) time: 16.1980 data: 15.4426 max mem: 8030 Epoch: [3] [60/70] eta: 0:02:36 lr: 0.005000 loss: 0.4335 (0.4263) loss_classifier: 0.1216 (0.1122) loss_box_reg: 0.2724 (0.2607) loss_objectness: 0.0166 (0.0218) loss_rpn_box_reg: 0.0254 (0.0316) time: 15.6932 data: 14.9200 max mem: 8030 Epoch: [4] [ 0/70] eta: 0:17:23 lr: 0.000500 loss: 0.2714 (0.2714) loss_classifier: 0.0745 (0.0745) loss_box_reg: 0.1744 (0.1744) loss_objectness: 0.0083 (0.0083) loss_rpn_box_reg: 0.0143 (0.0143) time: 14.9037 data: 14.0976 max mem: 8030 Epoch: [4] [10/70] eta: 0:16:11 lr: 0.000500 loss: 0.3816 (0.3980) loss_classifier: 0.0908 (0.0998) loss_box_reg: 0.2546 (0.2431) loss_objectness: 0.0169 (0.0239) loss_rpn_box_reg: 0.0270 (0.0312) time: 16.1879 data: 15.4277 max mem: 8030 Epoch: [4] [20/70] eta: 0:12:42 lr: 0.000500 loss: 0.3707 (0.3828) loss_classifier: 0.0908 (0.0974) loss_box_reg: 0.2391 (0.2354) loss_objectness: 0.0169 (0.0218) loss_rpn_box_reg: 0.0238 (0.0282) time: 15.2707 data: 14.5134 max mem: 8030 Epoch: [4] [30/70] eta: 0:10:24 lr: 0.000500 loss: 0.3507 (0.3781) loss_classifier: 0.0863 (0.0956) loss_box_reg: 0.2309 (0.2349) loss_objectness: 0.0154 (0.0200) loss_rpn_box_reg: 0.0227 (0.0275) time: 15.2926 data: 14.5223 max mem: 8030 Epoch: [4] [40/70] eta: 0:07:36 lr: 0.000500 loss: 0.3743 (0.3752) loss_classifier: 0.0882 (0.0954) loss_box_reg: 0.2319 (0.2343) loss_objectness: 0.0137 (0.0193) loss_rpn_box_reg: 0.0244 (0.0263) time: 15.1888 data: 14.4141 max mem: 8030 Epoch: [4] [50/70] eta: 0:05:02 lr: 0.000500 loss: 0.3743 (0.3796) loss_classifier: 0.0979 (0.0981) loss_box_reg: 0.2323 (0.2360) loss_objectness: 0.0175 (0.0188) loss_rpn_box_reg: 0.0243 (0.0266) time: 14.4223 data: 13.6610 max mem: 8030 Epoch: [4] [60/70] eta: 0:02:32 lr: 0.000500 loss: 0.3447 (0.3825) loss_classifier: 0.0996 (0.0982) loss_box_reg: 0.2239 (0.2375) loss_objectness: 0.0157 (0.0189) loss_rpn_box_reg: 0.0249 (0.0278) time: 15.3562 data: 14.6008 max mem: 8030 Epoch: [4] [69/70] eta: 0:00:15 lr: 0.000500 loss: 0.3327 (0.3821) loss_classifier: 0.0913 (0.0981) loss_box_reg: 0.2239 (0.2368) loss_objectness: 0.0147 (0.0188) loss_rpn_box_reg: 0.0234 (0.0283) time: 14.8779 data: 14.1158 max mem: 8030 Epoch: [4] Total time: 0:17:36 (15.0895 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:49 model_time: 0.3861 (0.3861) evaluator_time: 0.9473 (0.9473) time: 2.7638 data: 1.4287 max mem: 8030
Epoch: 15%|█▌ | 3/20 [1:15:09<5:21:39, 1135.27s/epoch, Train Loss=0.382, Test Loss=0.581, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3758 (0.3724) evaluator_time: 0.2049 (0.2651) time: 1.7110 data: 1.0716 max mem: 8030 Test: Total time: 0:00:30 (1.7111 s / it) Averaged stats: model_time: 0.3758 (0.3724) evaluator_time: 0.2049 (0.2651) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.581 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.657 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.002 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.650 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.020 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.191 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.649 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.129 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.650
Epoch: 20%|██ | 4/20 [1:15:10<4:59:44, 1124.02s/epoch, Train Loss=0.382, Test Loss=0.581, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [5] [ 0/70] eta: 0:21:06 lr: 0.000500 loss: 0.3157 (0.3157) loss_classifier: 0.0554 (0.0554) loss_box_reg: 0.2132 (0.2132) loss_objectness: 0.0195 (0.0195) loss_rpn_box_reg: 0.0276 (0.0276) time: 18.0962 data: 17.4030 max mem: 8030 Epoch: [5] [10/70] eta: 0:15:39 lr: 0.000500 loss: 0.3820 (0.3963) loss_classifier: 0.0997 (0.1030) loss_box_reg: 0.2300 (0.2440) loss_objectness: 0.0156 (0.0168) loss_rpn_box_reg: 0.0295 (0.0325) time: 15.6614 data: 14.9033 max mem: 8030 Epoch: [5] [20/70] eta: 0:12:58 lr: 0.000500 loss: 0.3562 (0.3868) loss_classifier: 0.0962 (0.0996) loss_box_reg: 0.2300 (0.2395) loss_objectness: 0.0113 (0.0154) loss_rpn_box_reg: 0.0249 (0.0323) time: 15.4366 data: 14.6783 max mem: 8030 Epoch: [5] [30/70] eta: 0:10:21 lr: 0.000500 loss: 0.3514 (0.3844) loss_classifier: 0.0855 (0.0967) loss_box_reg: 0.2314 (0.2411) loss_objectness: 0.0139 (0.0158) loss_rpn_box_reg: 0.0253 (0.0308) time: 15.4527 data: 14.6870 max mem: 8030 Epoch: [5] [40/70] eta: 0:07:38 lr: 0.000500 loss: 0.3529 (0.3765) loss_classifier: 0.0843 (0.0978) loss_box_reg: 0.2206 (0.2341) loss_objectness: 0.0144 (0.0156) loss_rpn_box_reg: 0.0258 (0.0290) time: 14.9956 data: 14.2181 max mem: 8030 Epoch: [5] [50/70] eta: 0:05:02 lr: 0.000500 loss: 0.3352 (0.3628) loss_classifier: 0.0804 (0.0929) loss_box_reg: 0.1941 (0.2278) loss_objectness: 0.0100 (0.0152) loss_rpn_box_reg: 0.0190 (0.0269) time: 14.4480 data: 13.6721 max mem: 8030 Epoch: [5] [60/70] eta: 0:02:31 lr: 0.000500 loss: 0.3147 (0.3629) loss_classifier: 0.0804 (0.0927) loss_box_reg: 0.2112 (0.2278) loss_objectness: 0.0086 (0.0153) loss_rpn_box_reg: 0.0190 (0.0270) time: 14.7789 data: 14.0017 max mem: 8030 Epoch: [5] [69/70] eta: 0:00:15 lr: 0.000500 loss: 0.3554 (0.3694) loss_classifier: 0.0926 (0.0943) loss_box_reg: 0.2361 (0.2309) loss_objectness: 0.0132 (0.0165) loss_rpn_box_reg: 0.0243 (0.0277) time: 15.0526 data: 14.2949 max mem: 8030 Epoch: [5] Total time: 0:17:39 (15.1349 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:48 model_time: 0.3762 (0.3762) evaluator_time: 0.9318 (0.9318) time: 2.6958 data: 1.3859 max mem: 8030
Epoch: 20%|██ | 4/20 [1:33:39<4:59:44, 1124.02s/epoch, Train Loss=0.369, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3762 (0.3722) evaluator_time: 0.1945 (0.2562) time: 1.6719 data: 1.0416 max mem: 8030 Test: Total time: 0:00:30 (1.6720 s / it) Averaged stats: model_time: 0.3762 (0.3722) evaluator_time: 0.1945 (0.2562) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.586 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.661 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 25%|██▌ | 5/20 [1:33:40<4:39:42, 1118.85s/epoch, Train Loss=0.369, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [6] [ 0/70] eta: 0:23:32 lr: 0.000500 loss: 0.3549 (0.3549) loss_classifier: 0.0845 (0.0845) loss_box_reg: 0.2291 (0.2291) loss_objectness: 0.0199 (0.0199) loss_rpn_box_reg: 0.0215 (0.0215) time: 20.1744 data: 19.4753 max mem: 8030 Epoch: [6] [10/70] eta: 0:16:23 lr: 0.000500 loss: 0.3399 (0.3449) loss_classifier: 0.0810 (0.0861) loss_box_reg: 0.2192 (0.2220) loss_objectness: 0.0106 (0.0131) loss_rpn_box_reg: 0.0224 (0.0236) time: 16.3859 data: 15.6274 max mem: 8030 Epoch: [6] [20/70] eta: 0:13:12 lr: 0.000500 loss: 0.3399 (0.3515) loss_classifier: 0.0911 (0.0917) loss_box_reg: 0.2156 (0.2206) loss_objectness: 0.0110 (0.0140) loss_rpn_box_reg: 0.0224 (0.0251) time: 15.6400 data: 14.8629 max mem: 8030 Epoch: [6] [30/70] eta: 0:10:22 lr: 0.000500 loss: 0.3397 (0.3389) loss_classifier: 0.0909 (0.0900) loss_box_reg: 0.1985 (0.2120) loss_objectness: 0.0128 (0.0132) loss_rpn_box_reg: 0.0186 (0.0238) time: 15.1102 data: 14.3225 max mem: 8030 Epoch: [6] [40/70] eta: 0:07:45 lr: 0.000500 loss: 0.3498 (0.3517) loss_classifier: 0.0855 (0.0914) loss_box_reg: 0.2202 (0.2220) loss_objectness: 0.0102 (0.0134) loss_rpn_box_reg: 0.0239 (0.0249) time: 15.1516 data: 14.3647 max mem: 8030 Epoch: [6] [50/70] eta: 0:05:12 lr: 0.000500 loss: 0.3851 (0.3588) loss_classifier: 0.0889 (0.0915) loss_box_reg: 0.2404 (0.2275) loss_objectness: 0.0122 (0.0140) loss_rpn_box_reg: 0.0258 (0.0258) time: 15.7605 data: 14.9888 max mem: 8030 Epoch: [6] [60/70] eta: 0:02:36 lr: 0.000500 loss: 0.3852 (0.3647) loss_classifier: 0.0973 (0.0930) loss_box_reg: 0.2402 (0.2297) loss_objectness: 0.0145 (0.0146) loss_rpn_box_reg: 0.0273 (0.0273) time: 15.8815 data: 15.1141 max mem: 8030 Epoch: [6] [69/70] eta: 0:00:15 lr: 0.000500 loss: 0.3888 (0.3663) loss_classifier: 0.0973 (0.0935) loss_box_reg: 0.2462 (0.2310) loss_objectness: 0.0145 (0.0145) loss_rpn_box_reg: 0.0276 (0.0273) time: 15.2576 data: 14.4892 max mem: 8030 Epoch: [6] Total time: 0:18:06 (15.5219 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:55 model_time: 0.3893 (0.3893) evaluator_time: 0.9429 (0.9429) time: 3.0642 data: 1.7303 max mem: 8030
Epoch: 25%|██▌ | 5/20 [1:52:38<4:39:42, 1118.85s/epoch, Train Loss=0.366, Test Loss=0.585, Train Accuracy=0.000, Test Accuracy=0.950]
Test: [17/18] eta: 0:00:01 model_time: 0.3762 (0.3683) evaluator_time: 0.1948 (0.2594) time: 1.7508 data: 1.1212 max mem: 8030 Test: Total time: 0:00:31 (1.7509 s / it) Averaged stats: model_time: 0.3762 (0.3683) evaluator_time: 0.1948 (0.2594) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.585 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.950 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.663 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.001 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.020 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.191 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.649 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.086 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651
Epoch: 25%|██▌ | 5/20 [1:52:39<5:37:57, 1351.84s/epoch, Train Loss=0.366, Test Loss=0.585, Train Accuracy=0.000, Test Accuracy=0.950]
Early stopping activated. Finished Training
Entrenamiento sin data-augmentation¶
En este punto, se realiza una prueba sin realizar data-augmentation:
In [39]:
def get_transforms(train):
transforms = []
if train:
pass
transforms.append(T.Resize(size=IMG_SIZE))
transforms.append(T.ToDtype(torch.float, scale=True))
transforms.append(T.ToPureTensor())
return T.Compose(transforms)
# Aplicamos las transformaciones.
train_dataset = TransformedPalmTreeDataset(train_dataset_raw, transforms=get_transforms(train=True))
test_dataset = TransformedPalmTreeDataset(test_dataset_raw, transforms=get_transforms(train=False))
save_checkpoint_path = os.path.join(RUTA_CARPETA_MODELS, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG + '_checkpoint.tar')
if LOAD_MODEL_FROM_CHECKPOINT and os.path.exists(save_checkpoint_path):
model, optimizer, lr_scheduler, checkpoint_data = get_fasterrcnn_resnet50_fpn_v2_model(len(LABELS.keys()), save_checkpoint_path, DEVICE)
else:
model, optimizer, lr_scheduler, checkpoint_data = get_fasterrcnn_resnet50_fpn_v2_model(len(LABELS.keys()), None, DEVICE)
EPOCHS = 20
BATCH_SIZE = 4
train_dataloader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True, collate_fn=lambda x: tuple(zip(*x)))
test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False, collate_fn=lambda x: tuple(zip(*x)))
train_losses, test_losses, train_accs, test_accs = fit(
model=model,
train_dataloader=train_dataloader,
test_dataloader=test_dataloader,
optimizer=optimizer,
criterion=None,
scheduler=lr_scheduler,
early_stopping=EarlyStopper(10, 0),
epochs=EPOCHS,
verbose=False,
device=DEVICE,
start_from_checkpoint_data=checkpoint_data,
base_dir=os.path.join(
RUTA_CARPETA_MODELS,
FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG),
model_id=FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG,
save_checkpoint=LOAD_MODEL_FROM_CHECKPOINT,
save_best_model = True,
save_results = True
)
Cargando el modelo desde el checkpoint: ./resources/models/fasterrcnn_resnet50_fpn_v2_sin_aug/fasterrcnn_resnet50_fpn_v2_sin_aug_checkpoint.tar Modelo cargado desde el checkpoint en la época 7. Traning from checkpoint at epoch 8
Epoch: 0%| | 0/13 [00:00<?, ?epoch/s]
Epoch: [8] [ 0/70] eta: 0:01:40 lr: 0.000050 loss: 0.2994 (0.2994) loss_classifier: 0.0674 (0.0674) loss_box_reg: 0.2092 (0.2092) loss_objectness: 0.0031 (0.0031) loss_rpn_box_reg: 0.0197 (0.0197) time: 1.4408 data: 0.6269 max mem: 8870 Epoch: [8] [10/70] eta: 0:01:31 lr: 0.000050 loss: 0.3304 (0.3396) loss_classifier: 0.0790 (0.0794) loss_box_reg: 0.2266 (0.2224) loss_objectness: 0.0100 (0.0127) loss_rpn_box_reg: 0.0224 (0.0251) time: 1.5230 data: 0.7672 max mem: 8870 Epoch: [8] [20/70] eta: 0:01:15 lr: 0.000050 loss: 0.3369 (0.3423) loss_classifier: 0.0872 (0.0845) loss_box_reg: 0.2266 (0.2206) loss_objectness: 0.0079 (0.0113) loss_rpn_box_reg: 0.0231 (0.0259) time: 1.5228 data: 0.7672 max mem: 8870 Epoch: [8] [30/70] eta: 0:01:00 lr: 0.000050 loss: 0.3567 (0.3502) loss_classifier: 0.0937 (0.0891) loss_box_reg: 0.2295 (0.2231) loss_objectness: 0.0084 (0.0116) loss_rpn_box_reg: 0.0262 (0.0264) time: 1.4925 data: 0.7328 max mem: 8870 Epoch: [8] [40/70] eta: 0:00:45 lr: 0.000050 loss: 0.3480 (0.3534) loss_classifier: 0.0981 (0.0910) loss_box_reg: 0.2202 (0.2235) loss_objectness: 0.0120 (0.0120) loss_rpn_box_reg: 0.0243 (0.0268) time: 1.5000 data: 0.7447 max mem: 8870 Epoch: [8] [50/70] eta: 0:00:30 lr: 0.000050 loss: 0.3467 (0.3509) loss_classifier: 0.0923 (0.0909) loss_box_reg: 0.2202 (0.2222) loss_objectness: 0.0107 (0.0117) loss_rpn_box_reg: 0.0215 (0.0260) time: 1.5053 data: 0.7600 max mem: 8870 Epoch: [8] [60/70] eta: 0:00:15 lr: 0.000050 loss: 0.3504 (0.3455) loss_classifier: 0.0918 (0.0887) loss_box_reg: 0.2310 (0.2200) loss_objectness: 0.0092 (0.0112) loss_rpn_box_reg: 0.0222 (0.0257) time: 1.5023 data: 0.7403 max mem: 8870 Epoch: [8] [69/70] eta: 0:00:01 lr: 0.000050 loss: 0.3598 (0.3468) loss_classifier: 0.0918 (0.0898) loss_box_reg: 0.2155 (0.2201) loss_objectness: 0.0090 (0.0115) loss_rpn_box_reg: 0.0220 (0.0254) time: 1.4737 data: 0.7056 max mem: 8870 Epoch: [8] Total time: 0:01:44 (1.4975 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3934 (0.3934) evaluator_time: 0.9314 (0.9314) time: 2.7986 data: 1.4720 max mem: 8870
Epoch: 0%| | 0/13 [02:35<?, ?epoch/s, Train Loss=0.347, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3763 (0.3740) evaluator_time: 0.1961 (0.2544) time: 1.6986 data: 1.0682 max mem: 8870 Test: Total time: 0:00:30 (1.6987 s / it) Averaged stats: model_time: 0.3763 (0.3740) evaluator_time: 0.1961 (0.2544) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.588 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.671 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.653 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.157 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.655
Epoch: 8%|▊ | 1/13 [02:36<31:15, 156.26s/epoch, Train Loss=0.347, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [9] [ 0/70] eta: 0:01:38 lr: 0.000050 loss: 0.2587 (0.2587) loss_classifier: 0.0728 (0.0728) loss_box_reg: 0.1634 (0.1634) loss_objectness: 0.0135 (0.0135) loss_rpn_box_reg: 0.0090 (0.0090) time: 1.4105 data: 0.6225 max mem: 8870 Epoch: [9] [10/70] eta: 0:01:31 lr: 0.000050 loss: 0.3870 (0.3617) loss_classifier: 0.1027 (0.0939) loss_box_reg: 0.2426 (0.2265) loss_objectness: 0.0135 (0.0133) loss_rpn_box_reg: 0.0284 (0.0279) time: 1.5197 data: 0.7438 max mem: 8870 Epoch: [9] [20/70] eta: 0:01:14 lr: 0.000050 loss: 0.3322 (0.3345) loss_classifier: 0.0751 (0.0868) loss_box_reg: 0.2172 (0.2120) loss_objectness: 0.0106 (0.0116) loss_rpn_box_reg: 0.0234 (0.0240) time: 1.4992 data: 0.7371 max mem: 8870 Epoch: [9] [30/70] eta: 0:00:59 lr: 0.000050 loss: 0.3096 (0.3379) loss_classifier: 0.0760 (0.0886) loss_box_reg: 0.1983 (0.2131) loss_objectness: 0.0081 (0.0121) loss_rpn_box_reg: 0.0198 (0.0242) time: 1.4855 data: 0.7227 max mem: 8870 Epoch: [9] [40/70] eta: 0:00:44 lr: 0.000050 loss: 0.3363 (0.3419) loss_classifier: 0.0885 (0.0890) loss_box_reg: 0.2160 (0.2160) loss_objectness: 0.0095 (0.0120) loss_rpn_box_reg: 0.0241 (0.0248) time: 1.4930 data: 0.7293 max mem: 8870 Epoch: [9] [50/70] eta: 0:00:29 lr: 0.000050 loss: 0.3523 (0.3426) loss_classifier: 0.0903 (0.0896) loss_box_reg: 0.2160 (0.2144) loss_objectness: 0.0113 (0.0127) loss_rpn_box_reg: 0.0280 (0.0259) time: 1.4838 data: 0.7327 max mem: 8870 Epoch: [9] [60/70] eta: 0:00:14 lr: 0.000050 loss: 0.3523 (0.3494) loss_classifier: 0.0935 (0.0913) loss_box_reg: 0.2306 (0.2194) loss_objectness: 0.0113 (0.0127) loss_rpn_box_reg: 0.0256 (0.0260) time: 1.4831 data: 0.7322 max mem: 8870 Epoch: [9] [69/70] eta: 0:00:01 lr: 0.000050 loss: 0.3244 (0.3447) loss_classifier: 0.0804 (0.0901) loss_box_reg: 0.2086 (0.2165) loss_objectness: 0.0095 (0.0124) loss_rpn_box_reg: 0.0227 (0.0256) time: 1.4836 data: 0.7308 max mem: 8870 Epoch: [9] Total time: 0:01:44 (1.4926 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:51 model_time: 0.3951 (0.3951) evaluator_time: 0.9898 (0.9898) time: 2.8577 data: 1.4710 max mem: 8870
Epoch: 8%|▊ | 1/13 [05:10<31:15, 156.26s/epoch, Train Loss=0.345, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3791 (0.3750) evaluator_time: 0.2029 (0.2608) time: 1.6965 data: 1.0588 max mem: 8870 Test: Total time: 0:00:30 (1.6966 s / it) Averaged stats: model_time: 0.3791 (0.3750) evaluator_time: 0.2029 (0.2608) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.588 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.670 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 15%|█▌ | 2/13 [05:11<28:32, 155.65s/epoch, Train Loss=0.345, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [10] [ 0/70] eta: 0:01:47 lr: 0.000005 loss: 0.2679 (0.2679) loss_classifier: 0.0770 (0.0770) loss_box_reg: 0.1590 (0.1590) loss_objectness: 0.0076 (0.0076) loss_rpn_box_reg: 0.0243 (0.0243) time: 1.5426 data: 0.7452 max mem: 8870 Epoch: [10] [10/70] eta: 0:01:28 lr: 0.000005 loss: 0.3374 (0.3447) loss_classifier: 0.0794 (0.0919) loss_box_reg: 0.2322 (0.2168) loss_objectness: 0.0091 (0.0100) loss_rpn_box_reg: 0.0263 (0.0261) time: 1.4817 data: 0.7246 max mem: 8870 Epoch: [10] [20/70] eta: 0:01:12 lr: 0.000005 loss: 0.3148 (0.3370) loss_classifier: 0.0849 (0.0909) loss_box_reg: 0.2067 (0.2106) loss_objectness: 0.0081 (0.0104) loss_rpn_box_reg: 0.0226 (0.0250) time: 1.4476 data: 0.6843 max mem: 8870 Epoch: [10] [30/70] eta: 0:00:58 lr: 0.000005 loss: 0.3496 (0.3491) loss_classifier: 0.0849 (0.0938) loss_box_reg: 0.2214 (0.2173) loss_objectness: 0.0103 (0.0115) loss_rpn_box_reg: 0.0264 (0.0266) time: 1.4506 data: 0.7005 max mem: 8870 Epoch: [10] [40/70] eta: 0:00:43 lr: 0.000005 loss: 0.3679 (0.3557) loss_classifier: 0.1007 (0.0945) loss_box_reg: 0.2374 (0.2230) loss_objectness: 0.0116 (0.0111) loss_rpn_box_reg: 0.0294 (0.0271) time: 1.4573 data: 0.7062 max mem: 8870 Epoch: [10] [50/70] eta: 0:00:29 lr: 0.000005 loss: 0.3650 (0.3495) loss_classifier: 0.0994 (0.0930) loss_box_reg: 0.2332 (0.2199) loss_objectness: 0.0088 (0.0108) loss_rpn_box_reg: 0.0219 (0.0259) time: 1.4505 data: 0.6813 max mem: 8870 Epoch: [10] [60/70] eta: 0:00:14 lr: 0.000005 loss: 0.3208 (0.3476) loss_classifier: 0.0767 (0.0914) loss_box_reg: 0.2044 (0.2191) loss_objectness: 0.0080 (0.0111) loss_rpn_box_reg: 0.0198 (0.0260) time: 1.4648 data: 0.7071 max mem: 8870 Epoch: [10] [69/70] eta: 0:00:01 lr: 0.000005 loss: 0.3208 (0.3413) loss_classifier: 0.0767 (0.0891) loss_box_reg: 0.2044 (0.2158) loss_objectness: 0.0082 (0.0111) loss_rpn_box_reg: 0.0256 (0.0253) time: 1.4687 data: 0.7093 max mem: 8870 Epoch: [10] Total time: 0:01:42 (1.4612 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:48 model_time: 0.3918 (0.3918) evaluator_time: 0.9290 (0.9290) time: 2.6700 data: 1.3474 max mem: 8870
Epoch: 15%|█▌ | 2/13 [07:41<28:32, 155.65s/epoch, Train Loss=0.341, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3781 (0.3743) evaluator_time: 0.1972 (0.2585) time: 1.6473 data: 1.0126 max mem: 8870 Test: Total time: 0:00:29 (1.6474 s / it) Averaged stats: model_time: 0.3781 (0.3743) evaluator_time: 0.1972 (0.2585) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.587 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.669 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.171 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 23%|██▎ | 3/13 [07:42<25:34, 153.48s/epoch, Train Loss=0.341, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [11] [ 0/70] eta: 0:01:55 lr: 0.000005 loss: 0.5099 (0.5099) loss_classifier: 0.1481 (0.1481) loss_box_reg: 0.3039 (0.3039) loss_objectness: 0.0116 (0.0116) loss_rpn_box_reg: 0.0463 (0.0463) time: 1.6438 data: 0.8500 max mem: 8870 Epoch: [11] [10/70] eta: 0:01:33 lr: 0.000005 loss: 0.3831 (0.3723) loss_classifier: 0.0955 (0.0951) loss_box_reg: 0.2453 (0.2389) loss_objectness: 0.0112 (0.0114) loss_rpn_box_reg: 0.0289 (0.0268) time: 1.5542 data: 0.8311 max mem: 8870 Epoch: [11] [20/70] eta: 0:01:16 lr: 0.000005 loss: 0.3591 (0.3667) loss_classifier: 0.0907 (0.0951) loss_box_reg: 0.2209 (0.2316) loss_objectness: 0.0111 (0.0132) loss_rpn_box_reg: 0.0241 (0.0267) time: 1.5265 data: 0.7866 max mem: 8870 Epoch: [11] [30/70] eta: 0:01:00 lr: 0.000005 loss: 0.3244 (0.3519) loss_classifier: 0.0869 (0.0924) loss_box_reg: 0.2159 (0.2224) loss_objectness: 0.0094 (0.0121) loss_rpn_box_reg: 0.0211 (0.0249) time: 1.5052 data: 0.7248 max mem: 8870 Epoch: [11] [40/70] eta: 0:00:45 lr: 0.000005 loss: 0.2995 (0.3401) loss_classifier: 0.0823 (0.0868) loss_box_reg: 0.2065 (0.2169) loss_objectness: 0.0095 (0.0120) loss_rpn_box_reg: 0.0203 (0.0244) time: 1.5249 data: 0.7380 max mem: 8871 Epoch: [11] [50/70] eta: 0:00:30 lr: 0.000005 loss: 0.2951 (0.3386) loss_classifier: 0.0792 (0.0869) loss_box_reg: 0.2013 (0.2158) loss_objectness: 0.0103 (0.0119) loss_rpn_box_reg: 0.0208 (0.0240) time: 1.5220 data: 0.7574 max mem: 8871 Epoch: [11] [60/70] eta: 0:00:15 lr: 0.000005 loss: 0.3643 (0.3480) loss_classifier: 0.0992 (0.0900) loss_box_reg: 0.2360 (0.2206) loss_objectness: 0.0101 (0.0122) loss_rpn_box_reg: 0.0236 (0.0251) time: 1.5235 data: 0.7702 max mem: 8871 Epoch: [11] [69/70] eta: 0:00:01 lr: 0.000005 loss: 0.3488 (0.3476) loss_classifier: 0.0895 (0.0900) loss_box_reg: 0.2105 (0.2196) loss_objectness: 0.0102 (0.0123) loss_rpn_box_reg: 0.0251 (0.0258) time: 1.4963 data: 0.7467 max mem: 8871 Epoch: [11] Total time: 0:01:46 (1.5171 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3932 (0.3932) evaluator_time: 0.9240 (0.9240) time: 2.8081 data: 1.4891 max mem: 8871
Epoch: 23%|██▎ | 3/13 [10:19<25:34, 153.48s/epoch, Train Loss=0.348, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3786 (0.3752) evaluator_time: 0.1958 (0.2584) time: 1.7207 data: 1.0852 max mem: 8871 Test: Total time: 0:00:30 (1.7208 s / it) Averaged stats: model_time: 0.3786 (0.3752) evaluator_time: 0.1958 (0.2584) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.588 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.670 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.129 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 31%|███ | 4/13 [10:20<23:16, 155.20s/epoch, Train Loss=0.348, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [12] [ 0/70] eta: 0:01:33 lr: 0.000005 loss: 0.2908 (0.2908) loss_classifier: 0.0856 (0.0856) loss_box_reg: 0.1752 (0.1752) loss_objectness: 0.0165 (0.0165) loss_rpn_box_reg: 0.0135 (0.0135) time: 1.3298 data: 0.5406 max mem: 8871 Epoch: [12] [10/70] eta: 0:01:30 lr: 0.000005 loss: 0.3086 (0.3305) loss_classifier: 0.0857 (0.0810) loss_box_reg: 0.1999 (0.2132) loss_objectness: 0.0097 (0.0117) loss_rpn_box_reg: 0.0200 (0.0247) time: 1.5043 data: 0.7602 max mem: 8871 Epoch: [12] [20/70] eta: 0:01:15 lr: 0.000005 loss: 0.3441 (0.3470) loss_classifier: 0.0876 (0.0898) loss_box_reg: 0.2200 (0.2197) loss_objectness: 0.0097 (0.0120) loss_rpn_box_reg: 0.0266 (0.0255) time: 1.5149 data: 0.7456 max mem: 8871 Epoch: [12] [30/70] eta: 0:01:00 lr: 0.000005 loss: 0.3607 (0.3539) loss_classifier: 0.0861 (0.0897) loss_box_reg: 0.2324 (0.2248) loss_objectness: 0.0105 (0.0121) loss_rpn_box_reg: 0.0277 (0.0274) time: 1.5204 data: 0.7446 max mem: 8871 Epoch: [12] [40/70] eta: 0:00:45 lr: 0.000005 loss: 0.3573 (0.3499) loss_classifier: 0.0861 (0.0916) loss_box_reg: 0.2190 (0.2207) loss_objectness: 0.0105 (0.0119) loss_rpn_box_reg: 0.0265 (0.0257) time: 1.5141 data: 0.7506 max mem: 8871 Epoch: [12] [50/70] eta: 0:00:30 lr: 0.000005 loss: 0.3405 (0.3456) loss_classifier: 0.0960 (0.0897) loss_box_reg: 0.2052 (0.2181) loss_objectness: 0.0094 (0.0118) loss_rpn_box_reg: 0.0226 (0.0259) time: 1.4986 data: 0.7298 max mem: 8871 Epoch: [12] [60/70] eta: 0:00:15 lr: 0.000005 loss: 0.3363 (0.3459) loss_classifier: 0.0912 (0.0899) loss_box_reg: 0.2157 (0.2183) loss_objectness: 0.0089 (0.0119) loss_rpn_box_reg: 0.0227 (0.0258) time: 1.5157 data: 0.7517 max mem: 8871 Epoch: [12] [69/70] eta: 0:00:01 lr: 0.000005 loss: 0.3426 (0.3454) loss_classifier: 0.0923 (0.0893) loss_box_reg: 0.2206 (0.2188) loss_objectness: 0.0102 (0.0117) loss_rpn_box_reg: 0.0247 (0.0256) time: 1.5309 data: 0.7792 max mem: 8871 Epoch: [12] Total time: 0:01:46 (1.5148 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3897 (0.3897) evaluator_time: 0.9351 (0.9351) time: 2.8219 data: 1.4954 max mem: 8871
Epoch: 31%|███ | 4/13 [12:57<23:16, 155.20s/epoch, Train Loss=0.345, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3773 (0.3742) evaluator_time: 0.1941 (0.2575) time: 1.7135 data: 1.0799 max mem: 8871 Test: Total time: 0:00:30 (1.7135 s / it) Averaged stats: model_time: 0.3773 (0.3742) evaluator_time: 0.1941 (0.2575) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.586 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.668 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.650 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 38%|███▊ | 5/13 [12:57<20:48, 156.06s/epoch, Train Loss=0.345, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [13] [ 0/70] eta: 0:01:45 lr: 0.000001 loss: 0.3433 (0.3433) loss_classifier: 0.0882 (0.0882) loss_box_reg: 0.2191 (0.2191) loss_objectness: 0.0105 (0.0105) loss_rpn_box_reg: 0.0255 (0.0255) time: 1.5110 data: 0.7203 max mem: 8871 Epoch: [13] [10/70] eta: 0:01:29 lr: 0.000001 loss: 0.3092 (0.3377) loss_classifier: 0.0854 (0.0903) loss_box_reg: 0.1964 (0.2100) loss_objectness: 0.0105 (0.0118) loss_rpn_box_reg: 0.0212 (0.0257) time: 1.4936 data: 0.7303 max mem: 8871 Epoch: [13] [20/70] eta: 0:01:14 lr: 0.000001 loss: 0.3092 (0.3486) loss_classifier: 0.0781 (0.0898) loss_box_reg: 0.1968 (0.2198) loss_objectness: 0.0100 (0.0121) loss_rpn_box_reg: 0.0212 (0.0268) time: 1.4977 data: 0.7473 max mem: 8871 Epoch: [13] [30/70] eta: 0:01:00 lr: 0.000001 loss: 0.3263 (0.3402) loss_classifier: 0.0768 (0.0863) loss_box_reg: 0.2145 (0.2156) loss_objectness: 0.0100 (0.0119) loss_rpn_box_reg: 0.0244 (0.0263) time: 1.5169 data: 0.7694 max mem: 8871 Epoch: [13] [40/70] eta: 0:00:45 lr: 0.000001 loss: 0.3263 (0.3433) loss_classifier: 0.0789 (0.0880) loss_box_reg: 0.2145 (0.2176) loss_objectness: 0.0106 (0.0123) loss_rpn_box_reg: 0.0244 (0.0254) time: 1.5229 data: 0.7667 max mem: 8871 Epoch: [13] [50/70] eta: 0:00:30 lr: 0.000001 loss: 0.3468 (0.3415) loss_classifier: 0.0921 (0.0885) loss_box_reg: 0.2125 (0.2167) loss_objectness: 0.0108 (0.0118) loss_rpn_box_reg: 0.0214 (0.0245) time: 1.4997 data: 0.7330 max mem: 8871 Epoch: [13] [60/70] eta: 0:00:15 lr: 0.000001 loss: 0.3179 (0.3422) loss_classifier: 0.0832 (0.0884) loss_box_reg: 0.2015 (0.2161) loss_objectness: 0.0114 (0.0127) loss_rpn_box_reg: 0.0214 (0.0251) time: 1.5108 data: 0.7417 max mem: 8871 Epoch: [13] [69/70] eta: 0:00:01 lr: 0.000001 loss: 0.3561 (0.3453) loss_classifier: 0.0874 (0.0890) loss_box_reg: 0.2259 (0.2185) loss_objectness: 0.0115 (0.0125) loss_rpn_box_reg: 0.0242 (0.0254) time: 1.5173 data: 0.7740 max mem: 8871 Epoch: [13] Total time: 0:01:45 (1.5077 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3890 (0.3890) evaluator_time: 0.9192 (0.9192) time: 2.7884 data: 1.4784 max mem: 8871
Epoch: 38%|███▊ | 5/13 [15:34<20:48, 156.06s/epoch, Train Loss=0.345, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.956]
Test: [17/18] eta: 0:00:01 model_time: 0.3779 (0.3751) evaluator_time: 0.1914 (0.2565) time: 1.7245 data: 1.0910 max mem: 8871 Test: Total time: 0:00:31 (1.7246 s / it) Averaged stats: model_time: 0.3779 (0.3751) evaluator_time: 0.1914 (0.2565) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.586 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.956 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.666 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.020 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 46%|████▌ | 6/13 [15:34<18:14, 156.43s/epoch, Train Loss=0.345, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.956]
Epoch: [14] [ 0/70] eta: 0:01:31 lr: 0.000001 loss: 0.3619 (0.3619) loss_classifier: 0.1108 (0.1108) loss_box_reg: 0.2242 (0.2242) loss_objectness: 0.0100 (0.0100) loss_rpn_box_reg: 0.0169 (0.0169) time: 1.3065 data: 0.5145 max mem: 8871 Epoch: [14] [10/70] eta: 0:01:30 lr: 0.000001 loss: 0.3114 (0.3298) loss_classifier: 0.0819 (0.0840) loss_box_reg: 0.2012 (0.2153) loss_objectness: 0.0089 (0.0113) loss_rpn_box_reg: 0.0171 (0.0191) time: 1.5075 data: 0.7318 max mem: 8871 Epoch: [14] [20/70] eta: 0:01:16 lr: 0.000001 loss: 0.3254 (0.3336) loss_classifier: 0.0820 (0.0868) loss_box_reg: 0.2156 (0.2143) loss_objectness: 0.0090 (0.0113) loss_rpn_box_reg: 0.0194 (0.0212) time: 1.5399 data: 0.7712 max mem: 8871 Epoch: [14] [30/70] eta: 0:01:00 lr: 0.000001 loss: 0.3351 (0.3382) loss_classifier: 0.0835 (0.0871) loss_box_reg: 0.2187 (0.2162) loss_objectness: 0.0095 (0.0110) loss_rpn_box_reg: 0.0231 (0.0240) time: 1.5245 data: 0.7548 max mem: 8871 Epoch: [14] [40/70] eta: 0:00:45 lr: 0.000001 loss: 0.3327 (0.3414) loss_classifier: 0.0871 (0.0874) loss_box_reg: 0.2159 (0.2176) loss_objectness: 0.0095 (0.0114) loss_rpn_box_reg: 0.0249 (0.0250) time: 1.5211 data: 0.7564 max mem: 8871 Epoch: [14] [50/70] eta: 0:00:30 lr: 0.000001 loss: 0.3327 (0.3394) loss_classifier: 0.0900 (0.0881) loss_box_reg: 0.2159 (0.2146) loss_objectness: 0.0095 (0.0113) loss_rpn_box_reg: 0.0249 (0.0253) time: 1.5364 data: 0.7663 max mem: 8871 Epoch: [14] [60/70] eta: 0:00:15 lr: 0.000001 loss: 0.3378 (0.3392) loss_classifier: 0.0879 (0.0880) loss_box_reg: 0.2215 (0.2147) loss_objectness: 0.0081 (0.0111) loss_rpn_box_reg: 0.0234 (0.0254) time: 1.5254 data: 0.7506 max mem: 8871 Epoch: [14] [69/70] eta: 0:00:01 lr: 0.000001 loss: 0.3491 (0.3435) loss_classifier: 0.0827 (0.0884) loss_box_reg: 0.2303 (0.2180) loss_objectness: 0.0091 (0.0115) loss_rpn_box_reg: 0.0258 (0.0257) time: 1.5317 data: 0.7875 max mem: 8871 Epoch: [14] Total time: 0:01:46 (1.5256 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3874 (0.3874) evaluator_time: 0.9244 (0.9244) time: 2.7978 data: 1.4843 max mem: 8871
Epoch: 46%|████▌ | 6/13 [18:13<18:14, 156.43s/epoch, Train Loss=0.343, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3783 (0.3749) evaluator_time: 0.1930 (0.2536) time: 1.7285 data: 1.0981 max mem: 8871 Test: Total time: 0:00:31 (1.7286 s / it) Averaged stats: model_time: 0.3783 (0.3749) evaluator_time: 0.1930 (0.2536) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.586 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.668 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 54%|█████▍ | 7/13 [18:13<15:43, 157.21s/epoch, Train Loss=0.343, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [15] [ 0/70] eta: 0:01:48 lr: 0.000001 loss: 0.3306 (0.3306) loss_classifier: 0.0763 (0.0763) loss_box_reg: 0.2297 (0.2297) loss_objectness: 0.0045 (0.0045) loss_rpn_box_reg: 0.0202 (0.0202) time: 1.5510 data: 0.7554 max mem: 8871 Epoch: [15] [10/70] eta: 0:01:34 lr: 0.000001 loss: 0.3840 (0.3767) loss_classifier: 0.0804 (0.0943) loss_box_reg: 0.2438 (0.2370) loss_objectness: 0.0136 (0.0140) loss_rpn_box_reg: 0.0269 (0.0315) time: 1.5747 data: 0.8087 max mem: 8871 Epoch: [15] [20/70] eta: 0:01:16 lr: 0.000001 loss: 0.3347 (0.3478) loss_classifier: 0.0817 (0.0891) loss_box_reg: 0.2273 (0.2202) loss_objectness: 0.0117 (0.0122) loss_rpn_box_reg: 0.0205 (0.0264) time: 1.5345 data: 0.7662 max mem: 8871 Epoch: [15] [30/70] eta: 0:01:01 lr: 0.000001 loss: 0.3075 (0.3453) loss_classifier: 0.0824 (0.0889) loss_box_reg: 0.2066 (0.2172) loss_objectness: 0.0097 (0.0129) loss_rpn_box_reg: 0.0201 (0.0264) time: 1.5190 data: 0.7502 max mem: 8871 Epoch: [15] [40/70] eta: 0:00:46 lr: 0.000001 loss: 0.3294 (0.3470) loss_classifier: 0.0842 (0.0908) loss_box_reg: 0.2127 (0.2178) loss_objectness: 0.0095 (0.0121) loss_rpn_box_reg: 0.0214 (0.0263) time: 1.5559 data: 0.7806 max mem: 8871 Epoch: [15] [50/70] eta: 0:00:30 lr: 0.000001 loss: 0.3090 (0.3409) loss_classifier: 0.0842 (0.0894) loss_box_reg: 0.2002 (0.2146) loss_objectness: 0.0082 (0.0113) loss_rpn_box_reg: 0.0223 (0.0256) time: 1.5162 data: 0.7302 max mem: 8871 Epoch: [15] [60/70] eta: 0:00:15 lr: 0.000001 loss: 0.3265 (0.3364) loss_classifier: 0.0846 (0.0883) loss_box_reg: 0.2074 (0.2123) loss_objectness: 0.0083 (0.0110) loss_rpn_box_reg: 0.0212 (0.0248) time: 1.4796 data: 0.7112 max mem: 8871 Epoch: [15] [69/70] eta: 0:00:01 lr: 0.000001 loss: 0.3521 (0.3421) loss_classifier: 0.0885 (0.0890) loss_box_reg: 0.2249 (0.2158) loss_objectness: 0.0098 (0.0117) loss_rpn_box_reg: 0.0255 (0.0255) time: 1.5015 data: 0.7568 max mem: 8871 Epoch: [15] Total time: 0:01:46 (1.5199 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3824 (0.3824) evaluator_time: 0.9304 (0.9304) time: 2.8024 data: 1.4879 max mem: 8871
Epoch: 54%|█████▍ | 7/13 [20:50<15:43, 157.21s/epoch, Train Loss=0.342, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3795 (0.3750) evaluator_time: 0.1967 (0.2558) time: 1.7161 data: 1.0834 max mem: 8871 Test: Total time: 0:00:30 (1.7162 s / it) Averaged stats: model_time: 0.3795 (0.3750) evaluator_time: 0.1967 (0.2558) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.586 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.667 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.020 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.129 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 62%|██████▏ | 8/13 [20:51<13:07, 157.42s/epoch, Train Loss=0.342, Test Loss=0.586, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [16] [ 0/70] eta: 0:01:42 lr: 0.000000 loss: 0.4488 (0.4488) loss_classifier: 0.1038 (0.1038) loss_box_reg: 0.2909 (0.2909) loss_objectness: 0.0161 (0.0161) loss_rpn_box_reg: 0.0379 (0.0379) time: 1.4672 data: 0.6637 max mem: 8871 Epoch: [16] [10/70] eta: 0:01:31 lr: 0.000000 loss: 0.3422 (0.3188) loss_classifier: 0.0819 (0.0799) loss_box_reg: 0.2155 (0.2048) loss_objectness: 0.0101 (0.0119) loss_rpn_box_reg: 0.0225 (0.0222) time: 1.5277 data: 0.7398 max mem: 8871 Epoch: [16] [20/70] eta: 0:01:17 lr: 0.000000 loss: 0.3119 (0.3235) loss_classifier: 0.0756 (0.0865) loss_box_reg: 0.2024 (0.2029) loss_objectness: 0.0098 (0.0113) loss_rpn_box_reg: 0.0187 (0.0228) time: 1.5582 data: 0.7767 max mem: 8871 Epoch: [16] [30/70] eta: 0:01:01 lr: 0.000000 loss: 0.3059 (0.3259) loss_classifier: 0.0791 (0.0852) loss_box_reg: 0.2024 (0.2063) loss_objectness: 0.0092 (0.0113) loss_rpn_box_reg: 0.0206 (0.0232) time: 1.5388 data: 0.7632 max mem: 8871 Epoch: [16] [40/70] eta: 0:00:45 lr: 0.000000 loss: 0.3412 (0.3365) loss_classifier: 0.0819 (0.0871) loss_box_reg: 0.2198 (0.2133) loss_objectness: 0.0099 (0.0117) loss_rpn_box_reg: 0.0231 (0.0244) time: 1.4884 data: 0.7133 max mem: 8871 Epoch: [16] [50/70] eta: 0:00:30 lr: 0.000000 loss: 0.3532 (0.3332) loss_classifier: 0.0819 (0.0861) loss_box_reg: 0.2237 (0.2119) loss_objectness: 0.0099 (0.0112) loss_rpn_box_reg: 0.0251 (0.0240) time: 1.5233 data: 0.7536 max mem: 8871 Epoch: [16] [60/70] eta: 0:00:15 lr: 0.000000 loss: 0.3523 (0.3382) loss_classifier: 0.0849 (0.0876) loss_box_reg: 0.2206 (0.2141) loss_objectness: 0.0099 (0.0113) loss_rpn_box_reg: 0.0228 (0.0253) time: 1.5466 data: 0.7884 max mem: 8871 Epoch: [16] [69/70] eta: 0:00:01 lr: 0.000000 loss: 0.3523 (0.3421) loss_classifier: 0.0836 (0.0885) loss_box_reg: 0.2261 (0.2166) loss_objectness: 0.0111 (0.0114) loss_rpn_box_reg: 0.0256 (0.0255) time: 1.4863 data: 0.7431 max mem: 8871 Epoch: [16] Total time: 0:01:46 (1.5195 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3858 (0.3858) evaluator_time: 0.9365 (0.9365) time: 2.8080 data: 1.4839 max mem: 8871
Epoch: 62%|██████▏ | 8/13 [23:28<13:07, 157.42s/epoch, Train Loss=0.342, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3786 (0.3747) evaluator_time: 0.1912 (0.2569) time: 1.7087 data: 1.0752 max mem: 8871 Test: Total time: 0:00:30 (1.7088 s / it) Averaged stats: model_time: 0.3786 (0.3747) evaluator_time: 0.1912 (0.2569) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.587 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.669 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 69%|██████▉ | 9/13 [23:29<10:30, 157.54s/epoch, Train Loss=0.342, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [17] [ 0/70] eta: 0:01:35 lr: 0.000000 loss: 0.2973 (0.2973) loss_classifier: 0.0738 (0.0738) loss_box_reg: 0.1960 (0.1960) loss_objectness: 0.0039 (0.0039) loss_rpn_box_reg: 0.0235 (0.0235) time: 1.3574 data: 0.6805 max mem: 8871 Epoch: [17] [10/70] eta: 0:01:29 lr: 0.000000 loss: 0.3502 (0.3682) loss_classifier: 0.0789 (0.0866) loss_box_reg: 0.2384 (0.2409) loss_objectness: 0.0072 (0.0100) loss_rpn_box_reg: 0.0271 (0.0306) time: 1.4979 data: 0.7532 max mem: 8871 Epoch: [17] [20/70] eta: 0:01:17 lr: 0.000000 loss: 0.3823 (0.3702) loss_classifier: 0.0885 (0.0908) loss_box_reg: 0.2377 (0.2362) loss_objectness: 0.0108 (0.0117) loss_rpn_box_reg: 0.0286 (0.0314) time: 1.5563 data: 0.8018 max mem: 8871 Epoch: [17] [30/70] eta: 0:01:01 lr: 0.000000 loss: 0.3507 (0.3530) loss_classifier: 0.0885 (0.0886) loss_box_reg: 0.2193 (0.2261) loss_objectness: 0.0099 (0.0109) loss_rpn_box_reg: 0.0241 (0.0274) time: 1.5430 data: 0.7827 max mem: 8872 Epoch: [17] [40/70] eta: 0:00:45 lr: 0.000000 loss: 0.3372 (0.3522) loss_classifier: 0.0885 (0.0897) loss_box_reg: 0.2099 (0.2243) loss_objectness: 0.0102 (0.0115) loss_rpn_box_reg: 0.0200 (0.0267) time: 1.4949 data: 0.7256 max mem: 8872 Epoch: [17] [50/70] eta: 0:00:30 lr: 0.000000 loss: 0.3385 (0.3493) loss_classifier: 0.0900 (0.0891) loss_box_reg: 0.2105 (0.2226) loss_objectness: 0.0102 (0.0116) loss_rpn_box_reg: 0.0247 (0.0260) time: 1.4963 data: 0.7275 max mem: 8872 Epoch: [17] [60/70] eta: 0:00:15 lr: 0.000000 loss: 0.3215 (0.3450) loss_classifier: 0.0834 (0.0883) loss_box_reg: 0.2049 (0.2194) loss_objectness: 0.0090 (0.0115) loss_rpn_box_reg: 0.0200 (0.0257) time: 1.4930 data: 0.7185 max mem: 8872 Epoch: [17] [69/70] eta: 0:00:01 lr: 0.000000 loss: 0.3293 (0.3426) loss_classifier: 0.0866 (0.0883) loss_box_reg: 0.2173 (0.2172) loss_objectness: 0.0092 (0.0116) loss_rpn_box_reg: 0.0209 (0.0255) time: 1.4855 data: 0.7083 max mem: 8872 Epoch: [17] Total time: 0:01:45 (1.5063 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3869 (0.3869) evaluator_time: 0.9268 (0.9268) time: 2.7971 data: 1.4815 max mem: 8872
Epoch: 69%|██████▉ | 9/13 [26:05<10:30, 157.54s/epoch, Train Loss=0.343, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3797 (0.3748) evaluator_time: 0.1995 (0.2570) time: 1.7098 data: 1.0760 max mem: 8872 Test: Total time: 0:00:30 (1.7099 s / it) Averaged stats: model_time: 0.3797 (0.3748) evaluator_time: 0.1995 (0.2570) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.588 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.672 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.157 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 77%|███████▋ | 10/13 [26:05<07:51, 157.17s/epoch, Train Loss=0.343, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [18] [ 0/70] eta: 0:01:49 lr: 0.000000 loss: 0.2253 (0.2253) loss_classifier: 0.0609 (0.0609) loss_box_reg: 0.1479 (0.1479) loss_objectness: 0.0060 (0.0060) loss_rpn_box_reg: 0.0105 (0.0105) time: 1.5623 data: 0.7678 max mem: 8872 Epoch: [18] [10/70] eta: 0:01:33 lr: 0.000000 loss: 0.3159 (0.3345) loss_classifier: 0.0759 (0.0859) loss_box_reg: 0.2072 (0.2153) loss_objectness: 0.0069 (0.0096) loss_rpn_box_reg: 0.0215 (0.0237) time: 1.5570 data: 0.7803 max mem: 8872 Epoch: [18] [20/70] eta: 0:01:17 lr: 0.000000 loss: 0.3325 (0.3414) loss_classifier: 0.0778 (0.0861) loss_box_reg: 0.2206 (0.2223) loss_objectness: 0.0074 (0.0089) loss_rpn_box_reg: 0.0235 (0.0241) time: 1.5436 data: 0.7681 max mem: 8872 Epoch: [18] [30/70] eta: 0:01:01 lr: 0.000000 loss: 0.3483 (0.3530) loss_classifier: 0.0893 (0.0902) loss_box_reg: 0.2330 (0.2284) loss_objectness: 0.0082 (0.0099) loss_rpn_box_reg: 0.0237 (0.0244) time: 1.5343 data: 0.7578 max mem: 8872 Epoch: [18] [40/70] eta: 0:00:46 lr: 0.000000 loss: 0.3463 (0.3418) loss_classifier: 0.0819 (0.0869) loss_box_reg: 0.2231 (0.2214) loss_objectness: 0.0093 (0.0100) loss_rpn_box_reg: 0.0225 (0.0234) time: 1.5309 data: 0.7496 max mem: 8872 Epoch: [18] [50/70] eta: 0:00:30 lr: 0.000000 loss: 0.3161 (0.3434) loss_classifier: 0.0851 (0.0891) loss_box_reg: 0.2045 (0.2196) loss_objectness: 0.0092 (0.0106) loss_rpn_box_reg: 0.0223 (0.0242) time: 1.5037 data: 0.7290 max mem: 8872 Epoch: [18] [60/70] eta: 0:00:15 lr: 0.000000 loss: 0.3315 (0.3448) loss_classifier: 0.0884 (0.0885) loss_box_reg: 0.2139 (0.2199) loss_objectness: 0.0103 (0.0109) loss_rpn_box_reg: 0.0240 (0.0255) time: 1.5189 data: 0.7721 max mem: 8872 Epoch: [18] [69/70] eta: 0:00:01 lr: 0.000000 loss: 0.3180 (0.3393) loss_classifier: 0.0795 (0.0871) loss_box_reg: 0.2078 (0.2161) loss_objectness: 0.0124 (0.0111) loss_rpn_box_reg: 0.0237 (0.0250) time: 1.5019 data: 0.7584 max mem: 8872 Epoch: [18] Total time: 0:01:46 (1.5230 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3903 (0.3903) evaluator_time: 0.9237 (0.9237) time: 2.7991 data: 1.4833 max mem: 8872
Epoch: 77%|███████▋ | 10/13 [28:42<07:51, 157.17s/epoch, Train Loss=0.339, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3788 (0.3753) evaluator_time: 0.1986 (0.2583) time: 1.7063 data: 1.0707 max mem: 8872 Test: Total time: 0:00:30 (1.7064 s / it) Averaged stats: model_time: 0.3788 (0.3753) evaluator_time: 0.1986 (0.2583) Accumulating evaluation results... DONE (t=0.05s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.587 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.666 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.143 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.653
Epoch: 85%|████████▍ | 11/13 [28:43<05:14, 157.23s/epoch, Train Loss=0.339, Test Loss=0.587, Train Accuracy=0.000, Test Accuracy=0.957]
Epoch: [19] [ 0/70] eta: 0:01:38 lr: 0.000000 loss: 0.3407 (0.3407) loss_classifier: 0.0735 (0.0735) loss_box_reg: 0.2373 (0.2373) loss_objectness: 0.0075 (0.0075) loss_rpn_box_reg: 0.0224 (0.0224) time: 1.4033 data: 0.7253 max mem: 8872 Epoch: [19] [10/70] eta: 0:01:31 lr: 0.000000 loss: 0.3407 (0.3511) loss_classifier: 0.0915 (0.0943) loss_box_reg: 0.2314 (0.2155) loss_objectness: 0.0126 (0.0135) loss_rpn_box_reg: 0.0224 (0.0278) time: 1.5289 data: 0.7433 max mem: 8872 Epoch: [19] [20/70] eta: 0:01:15 lr: 0.000000 loss: 0.3197 (0.3354) loss_classifier: 0.0845 (0.0894) loss_box_reg: 0.1985 (0.2082) loss_objectness: 0.0106 (0.0129) loss_rpn_box_reg: 0.0214 (0.0250) time: 1.5180 data: 0.7380 max mem: 8872 Epoch: [19] [30/70] eta: 0:00:59 lr: 0.000000 loss: 0.3197 (0.3373) loss_classifier: 0.0845 (0.0872) loss_box_reg: 0.1985 (0.2124) loss_objectness: 0.0106 (0.0128) loss_rpn_box_reg: 0.0221 (0.0249) time: 1.4733 data: 0.7322 max mem: 8872 Epoch: [19] [40/70] eta: 0:00:44 lr: 0.000000 loss: 0.3372 (0.3393) loss_classifier: 0.0903 (0.0880) loss_box_reg: 0.2076 (0.2124) loss_objectness: 0.0105 (0.0127) loss_rpn_box_reg: 0.0225 (0.0262) time: 1.4847 data: 0.7552 max mem: 8872 Epoch: [19] [50/70] eta: 0:00:30 lr: 0.000000 loss: 0.3301 (0.3382) loss_classifier: 0.0903 (0.0875) loss_box_reg: 0.2107 (0.2131) loss_objectness: 0.0082 (0.0118) loss_rpn_box_reg: 0.0223 (0.0257) time: 1.5269 data: 0.7686 max mem: 8872 Epoch: [19] [60/70] eta: 0:00:15 lr: 0.000000 loss: 0.3426 (0.3444) loss_classifier: 0.0894 (0.0885) loss_box_reg: 0.2241 (0.2182) loss_objectness: 0.0062 (0.0116) loss_rpn_box_reg: 0.0245 (0.0261) time: 1.5340 data: 0.7696 max mem: 8872 Epoch: [19] [69/70] eta: 0:00:01 lr: 0.000000 loss: 0.3244 (0.3426) loss_classifier: 0.0856 (0.0878) loss_box_reg: 0.2255 (0.2174) loss_objectness: 0.0097 (0.0116) loss_rpn_box_reg: 0.0216 (0.0259) time: 1.4695 data: 0.7149 max mem: 8872 Epoch: [19] Total time: 0:01:44 (1.4971 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:50 model_time: 0.3944 (0.3944) evaluator_time: 0.9351 (0.9351) time: 2.8310 data: 1.4998 max mem: 8872
Epoch: 85%|████████▍ | 11/13 [31:17<05:14, 157.23s/epoch, Train Loss=0.343, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Test: [17/18] eta: 0:00:01 model_time: 0.3790 (0.3752) evaluator_time: 0.1979 (0.2577) time: 1.7038 data: 1.0689 max mem: 8872 Test: Total time: 0:00:30 (1.7039 s / it) Averaged stats: model_time: 0.3790 (0.3752) evaluator_time: 0.1979 (0.2577) Accumulating evaluation results... DONE (t=0.07s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.588 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.958 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.672 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.193 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.653 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.157 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.654
Epoch: 92%|█████████▏| 12/13 [31:18<02:36, 156.70s/epoch, Train Loss=0.343, Test Loss=0.588, Train Accuracy=0.000, Test Accuracy=0.958]
Epoch: [20] [ 0/70] eta: 0:01:41 lr: 0.000000 loss: 0.3026 (0.3026) loss_classifier: 0.0798 (0.0798) loss_box_reg: 0.1961 (0.1961) loss_objectness: 0.0052 (0.0052) loss_rpn_box_reg: 0.0216 (0.0216) time: 1.4486 data: 0.7687 max mem: 8872 Epoch: [20] [10/70] eta: 0:01:28 lr: 0.000000 loss: 0.3601 (0.3791) loss_classifier: 0.1037 (0.1029) loss_box_reg: 0.2468 (0.2322) loss_objectness: 0.0171 (0.0155) loss_rpn_box_reg: 0.0233 (0.0285) time: 1.4833 data: 0.7282 max mem: 8872 Epoch: [20] [20/70] eta: 0:01:13 lr: 0.000000 loss: 0.3601 (0.3697) loss_classifier: 0.0932 (0.0971) loss_box_reg: 0.2348 (0.2312) loss_objectness: 0.0112 (0.0140) loss_rpn_box_reg: 0.0233 (0.0276) time: 1.4723 data: 0.7154 max mem: 8872 Epoch: [20] [30/70] eta: 0:00:59 lr: 0.000000 loss: 0.3570 (0.3620) loss_classifier: 0.0932 (0.0960) loss_box_reg: 0.2294 (0.2258) loss_objectness: 0.0097 (0.0133) loss_rpn_box_reg: 0.0245 (0.0269) time: 1.4959 data: 0.7383 max mem: 8872 Epoch: [20] [40/70] eta: 0:00:44 lr: 0.000000 loss: 0.3251 (0.3481) loss_classifier: 0.0839 (0.0907) loss_box_reg: 0.2174 (0.2203) loss_objectness: 0.0084 (0.0121) loss_rpn_box_reg: 0.0191 (0.0251) time: 1.4873 data: 0.7123 max mem: 8872 Epoch: [20] [50/70] eta: 0:00:29 lr: 0.000000 loss: 0.2837 (0.3446) loss_classifier: 0.0762 (0.0897) loss_box_reg: 0.1852 (0.2177) loss_objectness: 0.0088 (0.0123) loss_rpn_box_reg: 0.0186 (0.0249) time: 1.4540 data: 0.6847 max mem: 8872 Epoch: [20] [60/70] eta: 0:00:14 lr: 0.000000 loss: 0.3278 (0.3423) loss_classifier: 0.0805 (0.0889) loss_box_reg: 0.2088 (0.2168) loss_objectness: 0.0105 (0.0118) loss_rpn_box_reg: 0.0208 (0.0248) time: 1.4974 data: 0.7438 max mem: 8872 Epoch: [20] [69/70] eta: 0:00:01 lr: 0.000000 loss: 0.3401 (0.3434) loss_classifier: 0.0845 (0.0888) loss_box_reg: 0.2242 (0.2175) loss_objectness: 0.0082 (0.0117) loss_rpn_box_reg: 0.0230 (0.0254) time: 1.4569 data: 0.7066 max mem: 8872 Epoch: [20] Total time: 0:01:43 (1.4736 s / it) creating index... index created! Test: [ 0/18] eta: 0:00:49 model_time: 0.3931 (0.3931) evaluator_time: 0.9160 (0.9160) time: 2.7570 data: 1.4462 max mem: 8872
Epoch: 92%|█████████▏| 12/13 [33:50<02:36, 156.70s/epoch, Train Loss=0.343, Test Loss=0.585, Train Accuracy=0.000, Test Accuracy=0.957]
Test: [17/18] eta: 0:00:01 model_time: 0.3771 (0.3744) evaluator_time: 0.1945 (0.2552) time: 1.6764 data: 1.0449 max mem: 8872 Test: Total time: 0:00:30 (1.6765 s / it) Averaged stats: model_time: 0.3771 (0.3744) evaluator_time: 0.1945 (0.2552) Accumulating evaluation results... DONE (t=0.06s). IoU metric: bbox Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.585 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.957 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.662 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.003 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.021 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.192 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.651 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.157 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.652
Epoch: 100%|██████████| 13/13 [33:51<00:00, 156.28s/epoch, Train Loss=0.343, Test Loss=0.585, Train Accuracy=0.000, Test Accuracy=0.957]
Finished Training
Verificacion de resultados¶
🔮 Función auxiliar:
Imprime los resultados de un modelo.
In [40]:
def plot_results(train_losses: List[float], test_losses: List[float], train_accs: List[float], test_accs: List[float], title):
plt.figure(figsize=(10, 5))
plt.subplot(121)
plt.plot(train_losses, 'r')
plt.plot(test_losses, 'g')
plt.title('Loss')
plt.legend(['train', 'test'])
plt.xlabel('Epoch')
plt.ylabel('Loss')
plt.grid(True)
plt.subplot(122)
plt.plot(train_accs, 'r')
plt.plot(test_accs, 'g')
plt.title('Accuracy')
plt.legend(['train', 'test'])
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.grid(True)
plt.suptitle(title)
plt.show()
Verificamos los resultados:
In [41]:
# Cargar el archivo JSON
results_json_path = os.path.join(RUTA_CARPETA_MODELS, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID + '_results.json')
with open(results_json_path, 'r') as file:
data = json.load(file)
# Asignar variables
train_losses = data['train_losses']
test_losses = data['test_losses']
train_accs = data['train_accs']
test_accs = data['test_accs']
best_test_acc = data['best_test_acc']
plot_results(train_losses, test_losses, train_accs, test_accs, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID)
In [42]:
# Cargar el archivo JSON
results_json_path = os.path.join(RUTA_CARPETA_MODELS, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG + '_results.json')
with open(results_json_path, 'r') as file:
data = json.load(file)
# Asignar variables
train_losses = data['train_losses']
test_losses = data['test_losses']
train_accs = data['train_accs']
test_accs = data['test_accs']
best_test_acc = data['best_test_acc']
plot_results(train_losses, test_losses, train_accs, test_accs, FASTERRCNN_RESNET50_FPN_V2_MODEL_ID_SIN_AUG)
Prueba de inferencia¶
Podemos probar como fue nuestro modelo en una imagen:
In [43]:
image, _ = test_dataset[0]
plt.figure()
plt.imshow(image.permute(1, 2, 0))
Out[43]:
<matplotlib.image.AxesImage at 0x7e5812d1f790>
Y las predicciones son:
In [45]:
# Definir el umbral para mostrar las predicciones (por ejemplo, 0.5)
score_threshold = 0.5
# Realizar una inferencia en modo de evaluación
model.eval()
# Seleccionar una imagen y un target del conjunto de prueba
image, target = test_dataset[0] # El targets se utiliza para en nombre de la imagen.
model.to(DEVICE)
image = image.to(DEVICE)
# Realizar la predicción
with torch.no_grad():
# Pasar la imagen en una lista
prediction = model([image])[0]
# Filtrar las cajas, etiquetas y scores en función del umbral
predicted_boxes = prediction["boxes"].cpu()
predicted_scores = prediction["scores"].cpu()
predicted_labels = prediction["labels"].cpu()
image = image.cpu()
# Aplicar el umbral
# indices = predicted_scores > score_threshold
indices = nms(predicted_boxes, predicted_scores, iou_threshold=score_threshold)
filtered_boxes = predicted_boxes[indices]
filtered_labels = [
f"{list(LABELS.keys())[list(LABELS.values()).index(label)]} ({score:.2f})"
for label, score in zip(predicted_labels[indices], predicted_scores[indices])
]
# Configuración para visualización
config = {
"linewidth": 1,
"edgecolor": "r",
"color": "white",
"fontsize": 9,
"backgroundcolor": "red",
"putText": True,
"show_image_name": True,
"super_title": "Predicciones del Modelo",
"super_title_fontsize": 16,
"super_title_y": 0.90,
"from_dataset_voc": True,
}
# Verificar si "image_id" está en el target; si no, usar un nombre genérico
target_for_visualization = {
"boxes": filtered_boxes,
"labels": filtered_labels,
"img_path": target["img_path"]
}
# Visualizar la imagen con las predicciones
show_single_image(image, target_for_visualization, config)
También, se puede crear una función para utilizar en otro ambiente:
In [46]:
def show_model_prediction(
model: torch.nn.Module,
image_path: str,
config: dict,
get_transforms: Callable,
device: str,
score_threshold: float = 0.5,
) -> None:
"""
Muestra las predicciones de un modelo Faster R-CNN en una imagen dada.
Args:
model (torch.nn.Module): El modelo de detección de objetos.
image_path (str): Ruta de la imagen a procesar.
config (dict): Configuración para el estilo de las cajas y texto.
get_transforms (callable): Función que aplica las transformaciones necesarias a la imagen.
score_threshold (float): Umbral mínimo para mostrar las predicciones.
Returns:
None
"""
# Cargar la imagen
image = read_image(image_path)
# Aplicar las transformaciones a la imagen
transforms = get_transforms(train=False)
# image, _ = transforms(image, {})
image = transforms(image)
model.to(device)
image = image.to(device)
# Realizar la predicción
model.eval()
with torch.no_grad():
prediction = model([image])[0]
# Filtrar las cajas, etiquetas y scores en función del umbral
predicted_boxes = prediction["boxes"].cpu()
predicted_scores = prediction["scores"].cpu()
predicted_labels = prediction["labels"].cpu()
image = image.cpu()
# Aplico el umbral
indices = nms(predicted_boxes, predicted_scores, iou_threshold=score_threshold)
filtered_boxes = predicted_boxes[indices]
filtered_labels = [
f"{list(LABELS.keys())[list(LABELS.values()).index(label)]} ({score:.2f})"
for label, score in zip(predicted_labels[indices], predicted_scores[indices])
]
# Preparar el target para la visualización
target_for_visualization = {
"boxes": filtered_boxes,
"labels": filtered_labels,
"img_path": image_path,
}
# Visualizar la imagen con las predicciones
show_single_image(image, target_for_visualization, config)
Probamos predecir una imagen aleatoria:
In [47]:
image_path, _ = get_random_image_and_annotation(
RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_LABELS
)
config = {
"linewidth": 1,
"edgecolor": "r",
"color": "white",
"fontsize": 9,
"backgroundcolor": "red",
"putText": True,
"show_image_name": True,
"super_title": "Predicciones de para una imagen aleatoria",
"super_title_fontsize": 16,
"super_title_y": 0.85,
"from_dataset_voc": True,
}
show_model_prediction(model, image_path, config, get_transforms, DEVICE, 0.5)
YOLO¶
El algoritmo de una sola capa para detección de objetos en tiempo real YOLO o You Only Look Once, utiliza una única red neuronal convolucional para detectar objetos en imágenes. Existen múltiples versiones, siendo una de las últimas versiones la YOLOv8 (2023). Se utilizará esta versión y se hará una pequeña comparación, a modo de curiosidad, con la versión más reciente YOLOv11 (2024).
🤖 Consulta a IA (ChatGPT 4o): 🤖
❔ Prompt:
Podrías generar una introducción sobre el modelo YOLOv11? Que incluya sus componentes, sus ventajas y desventajas, así como también comparalo con otros modelos del estado del arte.
🔃 Respuesta:
Introducción al Modelo YOLOv11 El modelo YOLOv11 (You Only Look Once, versión 11) representa la última evolución en la familia de algoritmos YOLO para la detección de objetos en tiempo real. A lo largo de las versiones previas, YOLO ha sido ampliamente reconocido por su enfoque unificado y eficiente para detectar múltiples objetos en imágenes con alta velocidad, lo que lo ha convertido en uno de los métodos más populares en aplicaciones de visión por computadora, como sistemas de seguridad, automóviles autónomos y más. La versión YOLOv11 introduce varias mejoras en cuanto a arquitectura, rendimiento y precisión.
Componentes principales de YOLOv11 Backbone: La columna vertebral de YOLOv11 es una red neuronal convolucional (CNN) mejorada, diseñada específicamente para extraer características profundas de las imágenes de entrada. Esta nueva versión incluye arquitecturas más eficientes como una variación de CSPDarknet, que mejora la capacidad de aprendizaje con menos parámetros y una mejor gestión del ancho de banda de la red. Neck: En YOLOv11, el Neck ha sido optimizado para mejorar la capacidad de capturar características en diferentes escalas. Utiliza un módulo FPN (Feature Pyramid Network) y PAN (Path Aggregation Network) más avanzados, permitiendo una detección más precisa de objetos pequeños y difíciles de reconocer. Head: El Head del modelo, encargado de realizar las predicciones finales de clase y caja de contorno, ha sido refinado para ser más eficiente. Se ha mejorado el uso de las anclas, haciéndolo más adaptable a la detección de objetos de diversos tamaños. Mecanismos de atención: YOLOv11 incorpora nuevas técnicas de atención que permiten enfocar la red en las áreas más relevantes de la imagen, mejorando la precisión sin aumentar significativamente la carga computacional. Ventajas de YOLOv11 Desventajas de YOLOv11 Comparación con otros modelos del estado del arte Conclusión YOLOv11 marca un avance significativo en la familia YOLO, equilibrando velocidad y precisión para adaptarse a una amplia gama de aplicaciones de visión por computadora. Aunque enfrenta una competencia considerable en el estado del arte de la detección de objetos, sigue siendo una opción sobresaliente para escenarios en los que el tiempo de inferencia rápido es crucial. Las mejoras en la arquitectura del modelo, junto con su capacidad de manejar diferentes tamaños de objetos y escenas complejas, lo posicionan como una herramienta versátil en el campo de la inteligencia artificial aplicada.
Figura 4: Arquitectura YOLOv10 para referencia. No se encontró disponible la arquitectura de YOLOv11 aún. Extraída de: https://blog.roboflow.com/what-is-yolov10
Preparar el dataset¶
El formato de una etiqueta de YOLOv8 es: Como el conjunto de imágenes ya se encuentra dividido, para mantener la coherencia entre los modelos, se decide utilizar la misma división existente.<class_id> <x_center> <y_center> <width> <height>. El dataset de imágenes que se utiliza contiene un par de archivos en formato .csv con los datos de las etiquetas en cada una de las imágenes; así mismo, éstas se encuentran divididas en dos conjuntos: train y test.
In [12]:
# Etiquetas:
LABELS = {
'Tree' : 0,
'Palm' : 1
}
NOMBRE_MODELO_YOLO = 'palmeras'
RUTA_CARPETA_YOLO_PROYECTOS = 'yolo'
RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN = os.path.join(RUTA_CARPETA_YOLO_PROYECTOS, YOLOV8_MODEL_ID, 'train')
RUTA_CARPETA_YOLOV8_PROYECTOS_VAL = os.path.join(RUTA_CARPETA_YOLO_PROYECTOS, YOLOV8_MODEL_ID, 'val')
RUTA_CARPETA_YOLOV11_PROYECTOS_TRAIN = os.path.join(RUTA_CARPETA_YOLO_PROYECTOS, YOLOV11_MODEL_ID, 'train')
RUTA_CARPETA_YOLOV11_PROYECTOS_VAL = os.path.join(RUTA_CARPETA_YOLO_PROYECTOS, YOLOV11_MODEL_ID, 'val')
#RUTA_CARPETA_YOLOV8_PROYECTOS_BEST_TRAINED = os.path.join(RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, NOMBRE_MODELO_YOLO, 'weights', 'best.pt')
In [ ]:
train_df = pd.read_csv(RUTA_CARPETA_KAGGLE_DATSET_YOLO_TRAIN_CSV)
test_df = pd.read_csv(RUTA_CARPETA_KAGGLE_DATSET_YOLO_TEST_CSV)
# Se imprime el head de los dataframes para verificar que se cargaron correctamente
print(train_df.head())
print(test_df.head())
# Veamos las clases disponibles
print('--- Clases disponibles ---')
print('En Train: ', train_df['class'].unique())
print('En Test: ', test_df['class'].unique())
filename width height class xmin ymin xmax ymax
0 ck2oh0xx2sthb0838g3pmr4ky.jpg 4000 3000 Palm 1147 1694 1337 1863
1 ck2oh0xx2sthb0838g3pmr4ky.jpg 4000 3000 Palm 1392 1725 1567 1918
2 ck2oh0xx2sthb0838g3pmr4ky.jpg 4000 3000 Palm 1645 1816 1756 1933
3 ck2oh0xx2sthb0838g3pmr4ky.jpg 4000 3000 Palm 1913 1793 2104 1966
4 ck2oh0xx2sthb0838g3pmr4ky.jpg 4000 3000 Palm 2270 1817 2488 2034
filename width height class xmin ymin xmax ymax
0 ck2jfuaiwocqi07256rg8xca1.jpg 4864 3648 Tree 1978 1084 2349 1414
1 ck2jfuaiwocqi07256rg8xca1.jpg 4864 3648 Tree 2295 1947 2727 2465
2 ck2jfuaiwocqi07256rg8xca1.jpg 4864 3648 Tree 1358 1765 1912 2357
3 ck2jfuaiwocqi07256rg8xca1.jpg 4864 3648 Tree 0 2479 380 2981
4 ck2jfuaiwocqi07256rg8xca1.jpg 4864 3648 Tree 1664 2740 2268 3264
--- Clases disponibles ---
En Train: ['Palm' 'Tree']
En Test: ['Tree' 'Palm']
Se define una función para convertir en formato YOLO los datos guardados en los archivos CSV a formato aceptado por YOLO (.txt).
In [ ]:
# Función para convertir las coordenadas de las cajas de anotación a formato YOLO
def convert_to_yolo_format(row):
"""
Convierte las coordenadas de las cajas de anotación a formato YOLO.
Args:
row (pd.Series): Fila del DataFrame con las anotaciones.
Returns:
tuple: Una tupla con los datos en formato YOLO.
"""
try:
image_width = row['width']
image_height = row['height']
x_center = (row['xmin'] + row['xmax']) / 2.0 / image_width
y_center = (row['ymin'] + row['ymax']) / 2.0 / image_height
width = (row['xmax'] - row['xmin']) / image_width
height = (row['ymax'] - row['ymin']) / image_height
class_id = LABELS.get(row['class'], None)
if class_id is None:
print(f"Clase no encontrada: <{row['class']}> en la fila <{row}>")
return None
return class_id, x_center, y_center, width, height
except KeyError as e:
print(f"Falta un dato en la fila: {e}")
return None
except TypeError:
print("Datos invalidos encontrados.")
return None
# Función para guardar las anotaciones en formato YOLO
def save_yolo_annotations(df, output_dir):
"""
Guarda las anotaciones en formato YOLO.
Args:
df (pd.DataFrame): DataFrame con las anotaciones.
output_dir (str): Directorio de salida.
Returns:
None
"""
if not os.path.exists(output_dir):
os.makedirs(output_dir)
for _, row in df.iterrows():
yolo_data = convert_to_yolo_format(row)
filename = os.path.splitext(row['filename'])[0] + '.txt'
output_path = os.path.join(output_dir, filename)
if yolo_data:
with open(output_path, 'a') as f:
f.write(f"{yolo_data[0]} {yolo_data[1]:.6f} {yolo_data[2]:.6f} {yolo_data[3]:.6f} {yolo_data[4]:.6f}\n")
In [ ]:
save_yolo_annotations(train_df, RUTA_CARPETA_KAGGLE_DATASET_YOLO_TRAIN)
save_yolo_annotations(test_df, RUTA_CARPETA_KAGGLE_DATASET_YOLO_TEST)
El modelo YOLO espera que las imágenes se encuentren en el mismo lugar que los archivos de texto con la información sobre la etiqueta. Se debe tener un archivo de texto con todas las etiquetas de las detecciones por cada una de las imágenes.
In [ ]:
# Compiamos las imágenes a las carpetas correspondientes (train y test)
for _, row in train_df.iterrows():
shutil.copy(os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TRAIN_DATA, row['filename']), RUTA_CARPETA_KAGGLE_DATASET_YOLO_TRAIN)
for _, row in test_df.iterrows():
shutil.copy(os.path.join(RUTA_CARPETA_KAGGLE_DATASET_VOC_TEST_DATA, row['filename']), RUTA_CARPETA_KAGGLE_DATASET_YOLO_TEST)
Entrenamiento con YOLOv8 y YOLOv11¶
In [ ]:
print(DEVICE)
cuda
Se entrenan dos modelos con una pequeña diferencia en los hiperparámetros: el tamaño de la imagen de entrada. YOLO automaticamente optimiza el Learning Rate y el Momentum dependiendo del tamaño del dataset, las épocas, las iteraciones y las entradas (es decir, la complejidad), además de elegir el optimizador que considera mejor. Existe además una función adicional de Ajuste de hiperparámetros, pero no se pudo ejecutar en el equipo.
In [14]:
# Parámetros
YOLO_IMG_SIZE_1 = 640
YOLO_IMG_SIZE_2 = 480
YOLO_EPOCHS_1 = 100
YOLO_EPOCHS_2 = 200
# Parámetros para el entrenamiento de YOLO
params_1 = {
"data": YOLOV8_CONFIG_YAML,
"imgsz": YOLO_IMG_SIZE_1,
"epochs": YOLO_EPOCHS_1,
"device": DEVICE,
"name": f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}_r0',
"patience" : 20
}
params_2 = {
"data": YOLOV8_CONFIG_YAML,
"imgsz": YOLO_IMG_SIZE_2,
"epochs": YOLO_EPOCHS_2,
"device": DEVICE,
"name": f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_2}_epochs{YOLO_EPOCHS_2}_r0',
"patience" : 20
}
Se utilizan los modelos "nano" en sus versiones YOLOv8 y YOLO 11. Ambos son modelos que están pre-entrenados, con pesos e hiperparámetros ya definidos. Estos modelos están entrenados en el dataset de COCO y es una buena base generalizable.
In [ ]:
# Se instancian los modelos YOLOv8 nano
model_yolov8_p1 = YOLO("yolov8n.pt")
model_yolov8_p2 = YOLO("yolov8n.pt")
# Se instancian los modelos YOLOv11 nano
model_yolov11_p1 = YOLO("yolo11n.pt")
model_yolov11_p2 = YOLO("yolo11n.pt")
In [ ]:
%tensorboard --logdir yolo/yolov8/train/palmeras_imgsz640_epochs10_lr0.001
In [ ]:
# Se entrena el modelo base
model_yolov8_p1.train(project=RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, **params_1)
Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) engine\trainer: task=detect, mode=train, model=yolov8n.pt, data=./resources\utils\yolov8.yaml, epochs=100, time=None, patience=20, batch=16, imgsz=640, save=True, save_period=-1, cache=False, device=cuda, workers=8, project=yolo\yolov8\train, name=palmeras_imgsz640_epochs100_r0, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=yolo\yolov8\train\palmeras_imgsz640_epochs100_r0 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True] 3 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 4 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True] 5 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 6 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1] 16 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 123648 ultralytics.nn.modules.block.C2f [192, 128, 1] 19 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 22 [15, 18, 21] 1 431062 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] Model summary: 249 layers, 2,690,598 parameters, 2,690,582 gradients, 6.9 GFLOPs Transferred 313/391 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir yolo\yolov8\train\palmeras_imgsz640_epochs100_r0', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks with YOLO11n... AMP: checks passed
train: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\train.cache... 279 images, 0 backgrounds, 0 corrupt: 100%|██████████| 279/279 [00:00<?, ?it/s] val: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\test.cache... 70 images, 0 backgrounds, 0 corrupt: 100%|██████████| 70/70 [00:00<?, ?it/s]
Plotting labels to yolo\yolov8\train\palmeras_imgsz640_epochs100_r0\labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 63 weight(decay=0.0), 70 weight(decay=0.0005), 69 bias(decay=0.0) TensorBoard: model graph visualization added Image sizes 640 train, 640 val Using 8 dataloader workers Logging results to yolo\yolov8\train\palmeras_imgsz640_epochs100_r0 Starting training for 100 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
1/100 3.41G 2.093 3.675 1.648 438 640: 100%|██████████| 18/18 [00:06<00:00, 2.97it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.0727 0.481 0.16 0.0684
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
2/100 3.85G 1.55 2.833 1.149 528 640: 100%|██████████| 18/18 [00:04<00:00, 3.89it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.78it/s]
all 70 2670 0.101 0.601 0.369 0.195
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
3/100 3.75G 1.569 1.959 1.134 435 640: 100%|██████████| 18/18 [00:04<00:00, 4.24it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.53it/s]
all 70 2670 0.0979 0.541 0.325 0.163
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
4/100 3.68G 1.506 1.442 1.143 262 640: 100%|██████████| 18/18 [00:04<00:00, 4.25it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.93it/s]
all 70 2670 0.114 0.593 0.345 0.171
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
5/100 3.33G 1.422 1.167 1.138 266 640: 100%|██████████| 18/18 [00:04<00:00, 4.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.36it/s]
all 70 2670 0.133 0.751 0.461 0.261
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
6/100 3.23G 1.414 1.1 1.122 204 640: 100%|██████████| 18/18 [00:03<00:00, 4.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.02it/s]
all 70 2670 0.629 0.382 0.515 0.291
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
7/100 3.36G 1.416 1.01 1.125 408 640: 100%|██████████| 18/18 [00:04<00:00, 4.05it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.36it/s]
all 70 2670 0.63 0.547 0.576 0.3
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
8/100 4.64G 1.367 0.9549 1.098 477 640: 100%|██████████| 18/18 [00:05<00:00, 3.50it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.35it/s]
all 70 2670 0.601 0.557 0.596 0.3
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
9/100 3.47G 1.346 0.925 1.096 449 640: 100%|██████████| 18/18 [00:03<00:00, 5.15it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.13it/s]
all 70 2670 0.61 0.586 0.619 0.33
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
10/100 4.63G 1.326 0.879 1.092 383 640: 100%|██████████| 18/18 [00:03<00:00, 4.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.19it/s]
all 70 2670 0.705 0.646 0.695 0.374
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
11/100 3.86G 1.324 0.853 1.096 427 640: 100%|██████████| 18/18 [00:03<00:00, 5.00it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.43it/s]
all 70 2670 0.638 0.626 0.669 0.357
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
12/100 3.71G 1.322 0.8324 1.088 274 640: 100%|██████████| 18/18 [00:03<00:00, 4.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.68it/s]
all 70 2670 0.75 0.641 0.693 0.378
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
13/100 3.7G 1.342 0.8314 1.093 532 640: 100%|██████████| 18/18 [00:03<00:00, 4.84it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.79it/s]
all 70 2670 0.668 0.698 0.715 0.391
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
14/100 3.7G 1.338 0.8033 1.084 287 640: 100%|██████████| 18/18 [00:03<00:00, 4.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.94it/s]
all 70 2670 0.696 0.679 0.704 0.389
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
15/100 3.08G 1.324 0.8228 1.084 386 640: 100%|██████████| 18/18 [00:03<00:00, 4.86it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.30it/s]
all 70 2670 0.742 0.706 0.744 0.409
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
16/100 3.97G 1.282 0.7809 1.077 537 640: 100%|██████████| 18/18 [00:03<00:00, 4.55it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.69it/s]
all 70 2670 0.713 0.72 0.736 0.395
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
17/100 3.71G 1.341 0.7713 1.09 366 640: 100%|██████████| 18/18 [00:04<00:00, 4.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.46it/s]
all 70 2670 0.645 0.58 0.638 0.344
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
18/100 3.82G 1.325 0.7754 1.088 476 640: 100%|██████████| 18/18 [00:03<00:00, 4.60it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:02<00:00, 1.25it/s]
all 70 2670 0.685 0.724 0.695 0.381
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
19/100 3.26G 1.316 0.7701 1.074 444 640: 100%|██████████| 18/18 [00:03<00:00, 4.97it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.39it/s]
all 70 2670 0.732 0.695 0.731 0.407
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
20/100 3.93G 1.274 0.7366 1.078 423 640: 100%|██████████| 18/18 [00:04<00:00, 4.49it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.73it/s]
all 70 2670 0.715 0.712 0.743 0.413
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
21/100 4.04G 1.283 0.7247 1.079 379 640: 100%|██████████| 18/18 [00:03<00:00, 5.11it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.31it/s]
all 70 2670 0.725 0.752 0.754 0.418
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
22/100 3.92G 1.268 0.7268 1.072 634 640: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.30it/s]
all 70 2670 0.734 0.75 0.76 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
23/100 4.26G 1.26 0.7065 1.061 520 640: 100%|██████████| 18/18 [00:03<00:00, 5.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.87it/s]
all 70 2670 0.729 0.767 0.772 0.422
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
24/100 4.53G 1.233 0.7116 1.063 351 640: 100%|██████████| 18/18 [00:03<00:00, 4.82it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.92it/s]
all 70 2670 0.716 0.722 0.735 0.416
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
25/100 3.17G 1.245 0.7084 1.07 454 640: 100%|██████████| 18/18 [00:03<00:00, 5.07it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.96it/s]
all 70 2670 0.765 0.793 0.818 0.447
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
26/100 4.12G 1.249 0.6985 1.05 441 640: 100%|██████████| 18/18 [00:03<00:00, 4.81it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.63it/s]
all 70 2670 0.772 0.759 0.785 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
27/100 3.34G 1.25 0.6992 1.059 538 640: 100%|██████████| 18/18 [00:03<00:00, 4.89it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.81it/s]
all 70 2670 0.699 0.793 0.759 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
28/100 4.26G 1.237 0.6898 1.065 373 640: 100%|██████████| 18/18 [00:04<00:00, 3.91it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.77it/s]
all 70 2670 0.72 0.769 0.768 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
29/100 3.47G 1.236 0.6788 1.052 387 640: 100%|██████████| 18/18 [00:03<00:00, 4.74it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:02<00:00, 1.23it/s]
all 70 2670 0.766 0.732 0.776 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
30/100 3.92G 1.25 0.6893 1.055 568 640: 100%|██████████| 18/18 [00:03<00:00, 4.57it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.07it/s]
all 70 2670 0.777 0.756 0.783 0.431
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
31/100 3.75G 1.234 0.6713 1.052 272 640: 100%|██████████| 18/18 [00:03<00:00, 4.67it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.07it/s]
all 70 2670 0.768 0.787 0.798 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
32/100 3.89G 1.228 0.6708 1.065 362 640: 100%|██████████| 18/18 [00:03<00:00, 4.84it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.44it/s]
all 70 2670 0.747 0.743 0.747 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
33/100 3.74G 1.211 0.6659 1.053 374 640: 100%|██████████| 18/18 [00:03<00:00, 4.66it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.35it/s]
all 70 2670 0.825 0.722 0.801 0.452
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
34/100 4.25G 1.257 0.6712 1.055 434 640: 100%|██████████| 18/18 [00:04<00:00, 4.42it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.88it/s]
all 70 2670 0.765 0.756 0.782 0.432
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
35/100 4.18G 1.241 0.6659 1.043 524 640: 100%|██████████| 18/18 [00:03<00:00, 4.79it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.58it/s]
all 70 2670 0.746 0.779 0.783 0.437
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
36/100 3.84G 1.235 0.6513 1.052 306 640: 100%|██████████| 18/18 [00:04<00:00, 4.25it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.01it/s]
all 70 2670 0.742 0.767 0.779 0.436
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
37/100 3.32G 1.236 0.6644 1.059 322 640: 100%|██████████| 18/18 [00:03<00:00, 4.65it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.11it/s]
all 70 2670 0.798 0.755 0.795 0.438
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
38/100 3.9G 1.24 0.6505 1.055 347 640: 100%|██████████| 18/18 [00:03<00:00, 4.92it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.10it/s]
all 70 2670 0.769 0.765 0.786 0.418
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
39/100 3.96G 1.227 0.6525 1.054 438 640: 100%|██████████| 18/18 [00:03<00:00, 4.87it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.03it/s]
all 70 2670 0.78 0.79 0.804 0.451
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
40/100 4.16G 1.211 0.6353 1.052 433 640: 100%|██████████| 18/18 [00:03<00:00, 4.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.01it/s]
all 70 2670 0.798 0.768 0.806 0.453
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
41/100 3.92G 1.204 0.6337 1.043 384 640: 100%|██████████| 18/18 [00:03<00:00, 4.68it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.67it/s]
all 70 2670 0.764 0.782 0.807 0.448
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
42/100 3.47G 1.229 0.6529 1.055 343 640: 100%|██████████| 18/18 [00:03<00:00, 4.74it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.90it/s]
all 70 2670 0.773 0.769 0.791 0.435
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
43/100 3.89G 1.22 0.6427 1.055 419 640: 100%|██████████| 18/18 [00:04<00:00, 4.29it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.82it/s]
all 70 2670 0.791 0.76 0.799 0.446
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
44/100 3.61G 1.215 0.632 1.041 620 640: 100%|██████████| 18/18 [00:03<00:00, 5.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.51it/s]
all 70 2670 0.75 0.771 0.776 0.441
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
45/100 3.41G 1.174 0.6115 1.034 497 640: 100%|██████████| 18/18 [00:03<00:00, 4.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.18it/s]
all 70 2670 0.776 0.774 0.793 0.444
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
46/100 3.64G 1.211 0.6287 1.045 539 640: 100%|██████████| 18/18 [00:04<00:00, 4.44it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.13it/s]
all 70 2670 0.773 0.792 0.802 0.451
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
47/100 3.07G 1.197 0.6211 1.046 293 640: 100%|██████████| 18/18 [00:04<00:00, 4.43it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.08it/s]
all 70 2670 0.774 0.798 0.799 0.447
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
48/100 4.37G 1.183 0.6181 1.042 352 640: 100%|██████████| 18/18 [00:08<00:00, 2.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.46it/s]
all 70 2670 0.764 0.771 0.784 0.443
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
49/100 3.45G 1.179 0.6119 1.041 541 640: 100%|██████████| 18/18 [00:03<00:00, 5.33it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.79it/s]
all 70 2670 0.79 0.774 0.811 0.449
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
50/100 4.07G 1.214 0.6321 1.047 407 640: 100%|██████████| 18/18 [00:03<00:00, 5.15it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.75it/s]
all 70 2670 0.783 0.777 0.797 0.44
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
51/100 3.39G 1.191 0.6088 1.046 496 640: 100%|██████████| 18/18 [00:03<00:00, 4.62it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.93it/s]
all 70 2670 0.767 0.784 0.789 0.442
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
52/100 3.9G 1.191 0.6043 1.035 465 640: 100%|██████████| 18/18 [00:03<00:00, 4.66it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.13it/s]
all 70 2670 0.765 0.78 0.791 0.441
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
53/100 3.95G 1.194 0.6015 1.041 553 640: 100%|██████████| 18/18 [00:03<00:00, 4.82it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.00it/s]
all 70 2670 0.763 0.765 0.777 0.435
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
54/100 4.12G 1.16 0.6032 1.036 449 640: 100%|██████████| 18/18 [00:03<00:00, 4.87it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.26it/s]
all 70 2670 0.774 0.765 0.791 0.443
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
55/100 3.61G 1.175 0.5983 1.038 262 640: 100%|██████████| 18/18 [00:03<00:00, 4.87it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.36it/s]
all 70 2670 0.786 0.789 0.806 0.45
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
56/100 3.68G 1.163 0.5845 1.025 475 640: 100%|██████████| 18/18 [00:04<00:00, 3.99it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.23it/s]
all 70 2670 0.797 0.796 0.813 0.445
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
57/100 4.1G 1.167 0.6046 1.044 252 640: 100%|██████████| 18/18 [00:03<00:00, 5.36it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.73it/s]
all 70 2670 0.796 0.785 0.814 0.447
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
58/100 4.23G 1.167 0.5943 1.031 550 640: 100%|██████████| 18/18 [00:03<00:00, 5.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.83it/s]
all 70 2670 0.801 0.757 0.801 0.442
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
59/100 4.71G 1.17 0.6012 1.033 315 640: 100%|██████████| 18/18 [00:07<00:00, 2.28it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.35it/s]
all 70 2670 0.748 0.802 0.786 0.44
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
60/100 3.49G 1.162 0.5909 1.036 474 640: 100%|██████████| 18/18 [00:03<00:00, 5.38it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.89it/s]
all 70 2670 0.762 0.791 0.782 0.449
EarlyStopping: Training stopped early as no improvement observed in last 20 epochs. Best results observed at epoch 40, best model saved as best.pt.
To update EarlyStopping(patience=20) pass a new patience value, i.e. `patience=300` or use `patience=0` to disable EarlyStopping.
60 epochs completed in 0.109 hours.
Optimizer stripped from yolo\yolov8\train\palmeras_imgsz640_epochs100_r0\weights\last.pt, 5.6MB
Optimizer stripped from yolo\yolov8\train\palmeras_imgsz640_epochs100_r0\weights\best.pt, 5.6MB
Validating yolo\yolov8\train\palmeras_imgsz640_epochs100_r0\weights\best.pt...
Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB)
Model summary (fused): 186 layers, 2,684,758 parameters, 0 gradients, 6.8 GFLOPs
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:03<00:00, 1.12s/it]
all 70 2670 0.798 0.768 0.806 0.453
Tree 43 325 0.668 0.6 0.644 0.303
Palm 66 2345 0.928 0.936 0.968 0.604
Speed: 1.0ms preprocess, 3.5ms inference, 0.0ms loss, 7.0ms postprocess per image
Results saved to yolo\yolov8\train\palmeras_imgsz640_epochs100_r0
Out[ ]:
ultralytics.utils.metrics.DetMetrics object with attributes:
ap_class_index: array([0, 1])
box: ultralytics.utils.metrics.Metric object
confusion_matrix: <ultralytics.utils.metrics.ConfusionMatrix object at 0x000002258AD1CAD0>
curves: ['Precision-Recall(B)', 'F1-Confidence(B)', 'Precision-Confidence(B)', 'Recall-Confidence(B)']
curves_results: [[array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 1, 1, 1, ..., 0.001345, 0.00067252, 0],
[ 1, 1, 1, ..., 0.028517, 0.014258, 0]]), 'Recall', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.0792, 0.0792, 0.080565, ..., 0, 0, 0],
[ 0.29021, 0.29021, 0.29508, ..., 0, 0, 0]]), 'Confidence', 'F1'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.041345, 0.041345, 0.042089, ..., 1, 1, 1],
[ 0.17008, 0.17008, 0.17344, ..., 1, 1, 1]]), 'Confidence', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.93846, 0.93846, 0.93846, ..., 0, 0, 0],
[ 0.98806, 0.98806, 0.98806, ..., 0, 0, 0]]), 'Confidence', 'Recall']]
fitness: 0.4886148220553541
keys: ['metrics/precision(B)', 'metrics/recall(B)', 'metrics/mAP50(B)', 'metrics/mAP50-95(B)']
maps: array([ 0.30274, 0.60396])
names: {0: 'Tree', 1: 'Palm'}
plot: True
results_dict: {'metrics/precision(B)': 0.7980907820130512, 'metrics/recall(B)': 0.7680170575692964, 'metrics/mAP50(B)': 0.8059971252024734, 'metrics/mAP50-95(B)': 0.4533501217056742, 'fitness': 0.4886148220553541}
save_dir: WindowsPath('yolo/yolov8/train/palmeras_imgsz640_epochs100_r0')
speed: {'preprocess': 0.9570871080671038, 'inference': 3.5302979605538503, 'loss': 0.0, 'postprocess': 7.015698296683175}
task: 'detect'
In [ ]:
# Se entrena el modelo base
model_yolov8_p2.train(project=RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, **params_2)
Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) engine\trainer: task=detect, mode=train, model=yolov8n.pt, data=./resources\utils\yolov8.yaml, epochs=200, time=None, patience=20, batch=16, imgsz=480, save=True, save_period=-1, cache=False, device=cuda, workers=8, project=yolo\yolov8\train, name=palmeras_imgsz480_epochs200_r0, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=yolo\yolov8\train\palmeras_imgsz480_epochs200_r0 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True] 3 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 4 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True] 5 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 6 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1] 16 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 123648 ultralytics.nn.modules.block.C2f [192, 128, 1] 19 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 22 [15, 18, 21] 1 431062 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] Model summary: 249 layers, 2,690,598 parameters, 2,690,582 gradients, 6.9 GFLOPs Transferred 313/391 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir yolo\yolov8\train\palmeras_imgsz480_epochs200_r0', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks with YOLO11n... AMP: checks passed
train: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\train.cache... 279 images, 0 backgrounds, 0 corrupt: 100%|██████████| 279/279 [00:00<?, ?it/s] val: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\test.cache... 70 images, 0 backgrounds, 0 corrupt: 100%|██████████| 70/70 [00:00<?, ?it/s]
Plotting labels to yolo\yolov8\train\palmeras_imgsz480_epochs200_r0\labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 63 weight(decay=0.0), 70 weight(decay=0.0005), 69 bias(decay=0.0) TensorBoard: model graph visualization added Image sizes 480 train, 480 val Using 8 dataloader workers Logging results to yolo\yolov8\train\palmeras_imgsz480_epochs200_r0 Starting training for 200 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
1/200 1.99G 2.362 3.779 1.665 439 480: 100%|██████████| 18/18 [00:04<00:00, 4.17it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.64it/s]
all 70 2670 0.0534 0.338 0.0429 0.0199
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
2/200 2.23G 1.642 3.137 1.153 527 480: 100%|██████████| 18/18 [00:04<00:00, 3.73it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.0898 0.536 0.272 0.117
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
3/200 2.18G 1.644 2.239 1.106 435 480: 100%|██████████| 18/18 [00:04<00:00, 4.24it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.15it/s]
all 70 2670 0.096 0.464 0.205 0.0755
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
4/200 2.14G 1.589 1.612 1.096 262 480: 100%|██████████| 18/18 [00:03<00:00, 4.65it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.04it/s]
all 70 2670 0.121 0.51 0.261 0.128
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
5/200 1.94G 1.514 1.269 1.086 265 480: 100%|██████████| 18/18 [00:03<00:00, 4.68it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.91it/s]
all 70 2670 0.789 0.413 0.408 0.219
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
6/200 1.88G 1.489 1.196 1.071 204 480: 100%|██████████| 18/18 [00:03<00:00, 4.96it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.51it/s]
all 70 2670 0.91 0.368 0.434 0.234
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
7/200 1.96G 1.449 1.059 1.063 408 480: 100%|██████████| 18/18 [00:04<00:00, 4.25it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.52it/s]
all 70 2670 0.421 0.407 0.446 0.246
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
8/200 2.68G 1.42 0.9878 1.049 477 480: 100%|██████████| 18/18 [00:04<00:00, 4.35it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.37it/s]
all 70 2670 0.399 0.492 0.483 0.265
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
9/200 2.02G 1.407 0.9554 1.062 449 480: 100%|██████████| 18/18 [00:03<00:00, 5.61it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.19it/s]
all 70 2670 0.603 0.5 0.53 0.3
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
10/200 2.67G 1.371 0.9196 1.045 383 480: 100%|██████████| 18/18 [00:02<00:00, 6.75it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.67it/s]
all 70 2670 0.57 0.54 0.555 0.319
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
11/200 2.24G 1.363 0.8758 1.048 426 480: 100%|██████████| 18/18 [00:04<00:00, 4.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.30it/s]
all 70 2670 0.686 0.568 0.61 0.327
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
12/200 2.16G 1.384 0.8511 1.05 274 480: 100%|██████████| 18/18 [00:03<00:00, 5.35it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.91it/s]
all 70 2670 0.591 0.635 0.636 0.337
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
13/200 2.15G 1.409 0.847 1.054 531 480: 100%|██████████| 18/18 [00:02<00:00, 7.01it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.45it/s]
all 70 2670 0.601 0.646 0.643 0.346
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
14/200 2.15G 1.38 0.8169 1.039 287 480: 100%|██████████| 18/18 [00:03<00:00, 5.07it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.35it/s]
all 70 2670 0.649 0.606 0.64 0.357
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
15/200 1.86G 1.381 0.8255 1.047 382 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.10it/s]
all 70 2670 0.703 0.635 0.669 0.36
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
16/200 2.3G 1.324 0.7819 1.041 540 480: 100%|██████████| 18/18 [00:03<00:00, 5.13it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.29it/s]
all 70 2670 0.664 0.677 0.686 0.362
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
17/200 2.16G 1.337 0.7503 1.03 366 480: 100%|██████████| 18/18 [00:02<00:00, 6.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.80it/s]
all 70 2670 0.675 0.649 0.689 0.388
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
18/200 2.23G 1.328 0.7551 1.036 476 480: 100%|██████████| 18/18 [00:03<00:00, 5.43it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.09it/s]
all 70 2670 0.71 0.695 0.714 0.389
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
19/200 1.91G 1.342 0.7688 1.028 444 480: 100%|██████████| 18/18 [00:03<00:00, 5.49it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.37it/s]
all 70 2670 0.672 0.621 0.669 0.37
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
20/200 2.28G 1.304 0.7295 1.032 423 480: 100%|██████████| 18/18 [00:03<00:00, 4.68it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.41it/s]
all 70 2670 0.671 0.636 0.654 0.364
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
21/200 2.35G 1.323 0.7223 1.037 379 480: 100%|██████████| 18/18 [00:02<00:00, 7.29it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.30it/s]
all 70 2670 0.748 0.695 0.729 0.395
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
22/200 2.27G 1.327 0.7268 1.033 632 480: 100%|██████████| 18/18 [00:03<00:00, 4.90it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.738 0.715 0.731 0.397
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
23/200 2.47G 1.314 0.7157 1.023 520 480: 100%|██████████| 18/18 [00:02<00:00, 7.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.30it/s]
all 70 2670 0.79 0.701 0.757 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
24/200 2.62G 1.268 0.7115 1.029 351 480: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.42it/s]
all 70 2670 0.746 0.69 0.717 0.389
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
25/200 1.86G 1.285 0.7129 1.033 455 480: 100%|██████████| 18/18 [00:03<00:00, 5.45it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.12it/s]
all 70 2670 0.779 0.709 0.749 0.404
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
26/200 2.39G 1.298 0.7082 1.019 440 480: 100%|██████████| 18/18 [00:02<00:00, 6.33it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.99it/s]
all 70 2670 0.736 0.651 0.699 0.366
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
27/200 1.96G 1.295 0.7095 1.023 538 480: 100%|██████████| 18/18 [00:03<00:00, 5.28it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.24it/s]
all 70 2670 0.74 0.696 0.726 0.391
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
28/200 2.46G 1.286 0.699 1.033 372 480: 100%|██████████| 18/18 [00:03<00:00, 5.15it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.66it/s]
all 70 2670 0.701 0.652 0.678 0.373
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
29/200 2.02G 1.293 0.6926 1.021 387 480: 100%|██████████| 18/18 [00:03<00:00, 5.73it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.64it/s]
all 70 2670 0.744 0.69 0.714 0.373
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
30/200 2.28G 1.288 0.6927 1.017 568 480: 100%|██████████| 18/18 [00:02<00:00, 6.26it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.68it/s]
all 70 2670 0.775 0.709 0.756 0.41
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
31/200 2.18G 1.271 0.67 1.016 272 480: 100%|██████████| 18/18 [00:03<00:00, 4.79it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.55it/s]
all 70 2670 0.767 0.706 0.748 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
32/200 2.25G 1.277 0.6763 1.032 363 480: 100%|██████████| 18/18 [00:03<00:00, 5.26it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.88it/s]
all 70 2670 0.768 0.709 0.739 0.399
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
33/200 2.17G 1.254 0.6679 1.02 375 480: 100%|██████████| 18/18 [00:03<00:00, 5.05it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.61it/s]
all 70 2670 0.763 0.732 0.77 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
34/200 2.46G 1.292 0.6827 1.016 434 480: 100%|██████████| 18/18 [00:02<00:00, 6.26it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.29it/s]
all 70 2670 0.765 0.725 0.757 0.407
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
35/200 2.42G 1.271 0.6714 1.009 523 480: 100%|██████████| 18/18 [00:04<00:00, 4.40it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.53it/s]
all 70 2670 0.694 0.684 0.704 0.393
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
36/200 2.23G 1.278 0.66 1.016 306 480: 100%|██████████| 18/18 [00:04<00:00, 4.24it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.45it/s]
all 70 2670 0.71 0.729 0.73 0.404
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
37/200 1.92G 1.277 0.6603 1.023 321 480: 100%|██████████| 18/18 [00:02<00:00, 6.56it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.41it/s]
all 70 2670 0.721 0.708 0.728 0.404
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
38/200 2.27G 1.28 0.6568 1.02 347 480: 100%|██████████| 18/18 [00:03<00:00, 5.96it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.51it/s]
all 70 2670 0.713 0.732 0.744 0.395
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
39/200 2.3G 1.272 0.6555 1.022 439 480: 100%|██████████| 18/18 [00:03<00:00, 5.20it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.51it/s]
all 70 2670 0.758 0.722 0.758 0.415
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
40/200 2.41G 1.251 0.6471 1.019 433 480: 100%|██████████| 18/18 [00:02<00:00, 6.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.41it/s]
all 70 2670 0.795 0.718 0.76 0.413
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
41/200 2.28G 1.249 0.6418 1.015 384 480: 100%|██████████| 18/18 [00:03<00:00, 5.07it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.49it/s]
all 70 2670 0.761 0.723 0.757 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
42/200 2.24G 1.268 0.6577 1.022 342 480: 100%|██████████| 18/18 [00:04<00:00, 4.37it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.26it/s]
all 70 2670 0.739 0.728 0.753 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
43/200 2.26G 1.251 0.6487 1.019 419 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.25it/s]
all 70 2670 0.702 0.72 0.73 0.41
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
44/200 2.1G 1.255 0.6384 1.003 619 480: 100%|██████████| 18/18 [00:03<00:00, 5.57it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.03it/s]
all 70 2670 0.76 0.699 0.746 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
45/200 1.99G 1.228 0.6281 1 497 480: 100%|██████████| 18/18 [00:03<00:00, 4.69it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.03it/s]
all 70 2670 0.728 0.748 0.75 0.415
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
46/200 2.12G 1.246 0.6377 1.008 537 480: 100%|██████████| 18/18 [00:02<00:00, 7.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.86it/s]
all 70 2670 0.757 0.709 0.743 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
47/200 1.8G 1.25 0.6334 1.014 293 480: 100%|██████████| 18/18 [00:03<00:00, 5.35it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.13it/s]
all 70 2670 0.718 0.742 0.747 0.407
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
48/200 2.54G 1.23 0.6318 1.01 353 480: 100%|██████████| 18/18 [00:04<00:00, 3.67it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.59it/s]
all 70 2670 0.734 0.736 0.754 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
49/200 2.01G 1.227 0.6209 1.011 540 480: 100%|██████████| 18/18 [00:02<00:00, 6.05it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.46it/s]
all 70 2670 0.725 0.735 0.76 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
50/200 2.36G 1.26 0.6413 1.013 407 480: 100%|██████████| 18/18 [00:03<00:00, 4.75it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.797 0.737 0.768 0.423
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
51/200 1.98G 1.243 0.6262 1.017 495 480: 100%|██████████| 18/18 [00:02<00:00, 7.07it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.74it/s]
all 70 2670 0.766 0.699 0.724 0.405
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
52/200 2.26G 1.246 0.6194 1.008 466 480: 100%|██████████| 18/18 [00:03<00:00, 4.81it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.06it/s]
all 70 2670 0.75 0.751 0.761 0.416
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
53/200 2.3G 1.244 0.6157 1.012 553 480: 100%|██████████| 18/18 [00:02<00:00, 6.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.27it/s]
all 70 2670 0.719 0.744 0.749 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
54/200 2.39G 1.213 0.6196 1.01 449 480: 100%|██████████| 18/18 [00:02<00:00, 6.41it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.57it/s]
all 70 2670 0.754 0.741 0.766 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
55/200 2.11G 1.221 0.6142 1.007 261 480: 100%|██████████| 18/18 [00:03<00:00, 5.40it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.71it/s]
all 70 2670 0.767 0.76 0.779 0.423
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
56/200 2.14G 1.211 0.6022 0.9943 475 480: 100%|██████████| 18/18 [00:03<00:00, 5.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.99it/s]
all 70 2670 0.802 0.739 0.778 0.426
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
57/200 2.38G 1.227 0.629 1.019 252 480: 100%|██████████| 18/18 [00:03<00:00, 4.59it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.13it/s]
all 70 2670 0.774 0.747 0.781 0.412
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
58/200 2.45G 1.229 0.6188 1.006 549 480: 100%|██████████| 18/18 [00:02<00:00, 6.41it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.27it/s]
all 70 2670 0.779 0.739 0.761 0.409
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
59/200 2.79G 1.234 0.6251 1.011 315 480: 100%|██████████| 18/18 [00:04<00:00, 4.47it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.50it/s]
all 70 2670 0.745 0.735 0.76 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
60/200 2.04G 1.224 0.6097 1.011 475 480: 100%|██████████| 18/18 [00:03<00:00, 5.40it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.79it/s]
all 70 2670 0.749 0.754 0.775 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
61/200 2.28G 1.234 0.6065 0.9964 547 480: 100%|██████████| 18/18 [00:03<00:00, 5.65it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.65it/s]
all 70 2670 0.762 0.76 0.779 0.414
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
62/200 2.37G 1.231 0.6064 1.012 360 480: 100%|██████████| 18/18 [00:02<00:00, 6.91it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.17it/s]
all 70 2670 0.775 0.75 0.761 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
63/200 2.33G 1.209 0.5939 0.9981 473 480: 100%|██████████| 18/18 [00:03<00:00, 4.75it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.68it/s]
all 70 2670 0.786 0.743 0.781 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
64/200 2.69G 1.225 0.5912 0.999 395 480: 100%|██████████| 18/18 [00:04<00:00, 4.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.08it/s]
all 70 2670 0.795 0.75 0.775 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
65/200 2G 1.202 0.5903 0.9968 511 480: 100%|██████████| 18/18 [00:03<00:00, 4.61it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.47it/s]
all 70 2670 0.753 0.746 0.765 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
66/200 2.05G 1.22 0.6013 1.006 411 480: 100%|██████████| 18/18 [00:03<00:00, 5.89it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.40it/s]
all 70 2670 0.761 0.776 0.792 0.449
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
67/200 2.21G 1.177 0.5806 1.004 449 480: 100%|██████████| 18/18 [00:02<00:00, 6.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.98it/s]
all 70 2670 0.749 0.72 0.75 0.423
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
68/200 2.7G 1.245 0.61 1.013 392 480: 100%|██████████| 18/18 [00:04<00:00, 4.13it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.37it/s]
all 70 2670 0.799 0.739 0.779 0.424
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
69/200 2.17G 1.217 0.596 1.006 425 480: 100%|██████████| 18/18 [00:02<00:00, 6.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.65it/s]
all 70 2670 0.792 0.752 0.783 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
70/200 2.17G 1.195 0.5828 1.005 306 480: 100%|██████████| 18/18 [00:03<00:00, 5.25it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.92it/s]
all 70 2670 0.776 0.729 0.763 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
71/200 1.99G 1.179 0.5714 0.9992 504 480: 100%|██████████| 18/18 [00:03<00:00, 5.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.782 0.747 0.762 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
72/200 2.36G 1.234 0.5986 1.019 321 480: 100%|██████████| 18/18 [00:04<00:00, 3.86it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.62it/s]
all 70 2670 0.762 0.766 0.78 0.424
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
73/200 2.34G 1.205 0.5815 0.9973 497 480: 100%|██████████| 18/18 [00:02<00:00, 7.04it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.68it/s]
all 70 2670 0.814 0.75 0.785 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
74/200 2.28G 1.167 0.568 0.986 548 480: 100%|██████████| 18/18 [00:03<00:00, 5.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.774 0.748 0.78 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
75/200 2.04G 1.182 0.5755 0.9858 667 480: 100%|██████████| 18/18 [00:03<00:00, 5.93it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.46it/s]
all 70 2670 0.767 0.747 0.772 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
76/200 2.2G 1.203 0.5771 0.9962 537 480: 100%|██████████| 18/18 [00:04<00:00, 4.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.38it/s]
all 70 2670 0.763 0.716 0.758 0.422
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
77/200 1.89G 1.186 0.5712 0.9981 468 480: 100%|██████████| 18/18 [00:03<00:00, 4.89it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.24it/s]
all 70 2670 0.806 0.741 0.779 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
78/200 2.47G 1.2 0.5785 0.9911 520 480: 100%|██████████| 18/18 [00:03<00:00, 5.01it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.52it/s]
all 70 2670 0.787 0.74 0.789 0.435
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
79/200 2.2G 1.177 0.5775 0.9944 649 480: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.15it/s]
all 70 2670 0.796 0.753 0.79 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
80/200 1.83G 1.204 0.5809 1.011 302 480: 100%|██████████| 18/18 [00:02<00:00, 6.13it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.97it/s]
all 70 2670 0.745 0.776 0.765 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
81/200 2.1G 1.177 0.5659 0.9925 372 480: 100%|██████████| 18/18 [00:03<00:00, 5.36it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.61it/s]
all 70 2670 0.777 0.726 0.764 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
82/200 2.63G 1.183 0.569 1.005 436 480: 100%|██████████| 18/18 [00:02<00:00, 6.25it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.18it/s]
all 70 2670 0.782 0.737 0.775 0.416
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
83/200 2.07G 1.185 0.5792 0.9933 445 480: 100%|██████████| 18/18 [00:02<00:00, 6.50it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.27it/s]
all 70 2670 0.773 0.711 0.745 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
84/200 1.77G 1.179 0.5753 0.9952 717 480: 100%|██████████| 18/18 [00:05<00:00, 3.37it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.71it/s]
all 70 2670 0.751 0.729 0.739 0.412
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
85/200 1.97G 1.189 0.5715 0.9913 493 480: 100%|██████████| 18/18 [00:03<00:00, 4.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.99it/s]
all 70 2670 0.79 0.759 0.783 0.429
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
86/200 2.33G 1.18 0.5739 0.9924 366 480: 100%|██████████| 18/18 [00:02<00:00, 6.66it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.11it/s]
all 70 2670 0.783 0.756 0.777 0.424
EarlyStopping: Training stopped early as no improvement observed in last 20 epochs. Best results observed at epoch 66, best model saved as best.pt.
To update EarlyStopping(patience=20) pass a new patience value, i.e. `patience=300` or use `patience=0` to disable EarlyStopping.
86 epochs completed in 0.134 hours.
Optimizer stripped from yolo\yolov8\train\palmeras_imgsz480_epochs200_r0\weights\last.pt, 5.6MB
Optimizer stripped from yolo\yolov8\train\palmeras_imgsz480_epochs200_r0\weights\best.pt, 5.6MB
Validating yolo\yolov8\train\palmeras_imgsz480_epochs200_r0\weights\best.pt...
Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB)
Model summary (fused): 186 layers, 2,684,758 parameters, 0 gradients, 6.8 GFLOPs
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:03<00:00, 1.06s/it]
all 70 2670 0.76 0.776 0.791 0.449
Tree 43 325 0.592 0.631 0.621 0.291
Palm 66 2345 0.929 0.922 0.962 0.608
Speed: 0.4ms preprocess, 2.2ms inference, 0.0ms loss, 5.8ms postprocess per image
Results saved to yolo\yolov8\train\palmeras_imgsz480_epochs200_r0
Out[ ]:
ultralytics.utils.metrics.DetMetrics object with attributes:
ap_class_index: array([0, 1])
box: ultralytics.utils.metrics.Metric object
confusion_matrix: <ultralytics.utils.metrics.ConfusionMatrix object at 0x000001DD1640BC10>
curves: ['Precision-Recall(B)', 'F1-Confidence(B)', 'Precision-Confidence(B)', 'Recall-Confidence(B)']
curves_results: [[array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 1, 1, 1, ..., 0.00057085, 0.00028542, 0],
[ 1, 1, 1, ..., 0.018603, 0.0093015, 0]]), 'Recall', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.051158, 0.051158, 0.055628, ..., 0, 0, 0],
[ 0.37713, 0.37713, 0.41357, ..., 0, 0, 0]]), 'Confidence', 'F1'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.02632, 0.02632, 0.028693, ..., 1, 1, 1],
[ 0.23379, 0.23379, 0.26249, ..., 1, 1, 1]]), 'Confidence', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.90769, 0.90769, 0.90769, ..., 0, 0, 0],
[ 0.97484, 0.97484, 0.97441, ..., 0, 0, 0]]), 'Confidence', 'Recall']]
fitness: 0.4833088152300575
keys: ['metrics/precision(B)', 'metrics/recall(B)', 'metrics/mAP50(B)', 'metrics/mAP50-95(B)']
maps: array([ 0.29059, 0.60755])
names: {0: 'Tree', 1: 'Palm'}
plot: True
results_dict: {'metrics/precision(B)': 0.7602786575389766, 'metrics/recall(B)': 0.7761522060029522, 'metrics/mAP50(B)': 0.7914198185930055, 'metrics/mAP50-95(B)': 0.44907425930084105, 'fitness': 0.4833088152300575}
save_dir: WindowsPath('yolo/yolov8/train/palmeras_imgsz480_epochs200_r0')
speed: {'preprocess': 0.38558074406215126, 'inference': 2.200007438659668, 'loss': 0.0, 'postprocess': 5.842811720711844}
task: 'detect'
In [ ]:
# Se entrena el modelo base
model_yolov11_p1.train(project=RUTA_CARPETA_YOLOV11_PROYECTOS_TRAIN, **params_1)
Initial GPU Usage | ID | GPU | MEM | ------------------ | 0 | 4% | 21% | GPU Usage after emptying the cache | ID | GPU | MEM | ------------------ | 0 | 4% | 21% | Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) engine\trainer: task=detect, mode=train, model=yolo11n.pt, data=./resources\utils\yolov8.yaml, epochs=100, time=None, patience=20, batch=16, imgsz=640, save=True, save_period=-1, cache=False, device=cuda, workers=8, project=yolo\yolov11\train, name=palmeras_imgsz640_epochs100_r0, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=yolo\yolov11\train\palmeras_imgsz640_epochs100_r0 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 6640 ultralytics.nn.modules.block.C3k2 [32, 64, 1, False, 0.25] 3 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 4 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 5 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 6 -1 1 87040 ultralytics.nn.modules.block.C3k2 [128, 128, 1, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 346112 ultralytics.nn.modules.block.C3k2 [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 249728 ultralytics.nn.modules.block.C2PSA [256, 256, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 13 -1 1 111296 ultralytics.nn.modules.block.C3k2 [384, 128, 1, False] 14 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 15 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 16 -1 1 32096 ultralytics.nn.modules.block.C3k2 [256, 64, 1, False] 17 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 18 [-1, 13] 1 0 ultralytics.nn.modules.conv.Concat [1] 19 -1 1 86720 ultralytics.nn.modules.block.C3k2 [192, 128, 1, False] 20 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 21 [-1, 10] 1 0 ultralytics.nn.modules.conv.Concat [1] 22 -1 1 378880 ultralytics.nn.modules.block.C3k2 [384, 256, 1, True] 23 [16, 19, 22] 1 431062 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] YOLO11n summary: 319 layers, 2,590,230 parameters, 2,590,214 gradients, 6.4 GFLOPs Transferred 448/499 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir yolo\yolov11\train\palmeras_imgsz640_epochs100_r0', view at http://localhost:6006/ Freezing layer 'model.23.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks with YOLO11n... AMP: checks passed
train: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\train.cache... 279 images, 0 backgrounds, 0 corrupt: 100%|██████████| 279/279 [00:00<?, ?it/s] val: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\test.cache... 70 images, 0 backgrounds, 0 corrupt: 100%|██████████| 70/70 [00:00<?, ?it/s]
Plotting labels to yolo\yolov11\train\palmeras_imgsz640_epochs100_r0\labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 81 weight(decay=0.0), 88 weight(decay=0.0005), 87 bias(decay=0.0) TensorBoard: model graph visualization added Image sizes 640 train, 640 val Using 8 dataloader workers Logging results to yolo\yolov11\train\palmeras_imgsz640_epochs100_r0 Starting training for 100 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
1/100 3.54G 2.156 3.576 1.681 438 640: 100%|██████████| 18/18 [00:05<00:00, 3.51it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.0811 0.548 0.224 0.109
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
2/100 4.03G 1.594 2.52 1.175 528 640: 100%|██████████| 18/18 [00:04<00:00, 4.45it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.24it/s]
all 70 2670 0.0819 0.543 0.346 0.181
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
3/100 3.92G 1.561 1.537 1.146 435 640: 100%|██████████| 18/18 [00:04<00:00, 4.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.46it/s]
all 70 2670 0.932 0.291 0.386 0.207
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
4/100 3.86G 1.439 1.18 1.139 262 640: 100%|██████████| 18/18 [00:04<00:00, 4.39it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.10it/s]
all 70 2670 0.944 0.408 0.502 0.284
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
5/100 3.5G 1.402 1.037 1.134 266 640: 100%|██████████| 18/18 [00:03<00:00, 4.58it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.32it/s]
all 70 2670 0.529 0.576 0.579 0.317
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
6/100 3.4G 1.395 1.015 1.111 204 640: 100%|██████████| 18/18 [00:03<00:00, 4.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.92it/s]
all 70 2670 0.538 0.523 0.532 0.286
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
7/100 3.54G 1.415 0.9629 1.122 408 640: 100%|██████████| 18/18 [00:03<00:00, 4.66it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.13it/s]
all 70 2670 0.513 0.592 0.576 0.328
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
8/100 4.81G 1.384 0.8925 1.106 477 640: 100%|██████████| 18/18 [00:03<00:00, 4.50it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.82it/s]
all 70 2670 0.569 0.5 0.541 0.303
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
9/100 3.64G 1.348 0.8806 1.103 449 640: 100%|██████████| 18/18 [00:03<00:00, 5.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.96it/s]
all 70 2670 0.63 0.627 0.622 0.353
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
10/100 4.8G 1.311 0.8499 1.092 383 640: 100%|██████████| 18/18 [00:08<00:00, 2.11it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.05it/s]
all 70 2670 0.693 0.669 0.689 0.389
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
11/100 4.03G 1.296 0.8175 1.086 427 640: 100%|██████████| 18/18 [00:03<00:00, 5.17it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.20it/s]
all 70 2670 0.653 0.668 0.665 0.359
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
12/100 3.88G 1.312 0.798 1.092 274 640: 100%|██████████| 18/18 [00:03<00:00, 4.84it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.29it/s]
all 70 2670 0.695 0.663 0.69 0.385
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
13/100 3.88G 1.346 0.802 1.096 532 640: 100%|██████████| 18/18 [00:03<00:00, 5.08it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.28it/s]
all 70 2670 0.706 0.675 0.701 0.38
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
14/100 3.88G 1.335 0.7829 1.09 287 640: 100%|██████████| 18/18 [00:03<00:00, 4.76it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.42it/s]
all 70 2670 0.698 0.635 0.693 0.38
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
15/100 3.25G 1.322 0.8042 1.092 386 640: 100%|██████████| 18/18 [00:03<00:00, 5.14it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.07it/s]
all 70 2670 0.731 0.671 0.721 0.403
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
16/100 4.14G 1.298 0.7653 1.088 537 640: 100%|██████████| 18/18 [00:03<00:00, 4.93it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.63it/s]
all 70 2670 0.723 0.714 0.739 0.402
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
17/100 3.82G 1.332 0.7499 1.092 366 640: 100%|██████████| 18/18 [00:03<00:00, 5.02it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.68it/s]
all 70 2670 0.75 0.703 0.743 0.4
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
18/100 3.99G 1.314 0.7607 1.084 476 640: 100%|██████████| 18/18 [00:03<00:00, 4.91it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.89it/s]
all 70 2670 0.691 0.732 0.723 0.395
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
19/100 3.44G 1.314 0.7444 1.085 444 640: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.37it/s]
all 70 2670 0.747 0.745 0.763 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
20/100 4.11G 1.276 0.7253 1.079 423 640: 100%|██████████| 18/18 [00:03<00:00, 4.55it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.16it/s]
all 70 2670 0.752 0.756 0.779 0.424
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
21/100 4.22G 1.29 0.7077 1.087 379 640: 100%|██████████| 18/18 [00:03<00:00, 5.13it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.19it/s]
all 70 2670 0.79 0.712 0.772 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
22/100 4.09G 1.276 0.7106 1.076 634 640: 100%|██████████| 18/18 [00:03<00:00, 4.85it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.66it/s]
all 70 2670 0.767 0.756 0.789 0.431
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
23/100 4.43G 1.275 0.6932 1.072 520 640: 100%|██████████| 18/18 [00:03<00:00, 4.94it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.06it/s]
all 70 2670 0.752 0.695 0.755 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
24/100 4.71G 1.236 0.6957 1.067 351 640: 100%|██████████| 18/18 [00:03<00:00, 4.90it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.03it/s]
all 70 2670 0.756 0.711 0.755 0.417
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
25/100 3.35G 1.238 0.699 1.072 454 640: 100%|██████████| 18/18 [00:03<00:00, 5.04it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.98it/s]
all 70 2670 0.772 0.757 0.792 0.426
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
26/100 4.29G 1.262 0.6855 1.06 441 640: 100%|██████████| 18/18 [00:03<00:00, 4.67it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.90it/s]
all 70 2670 0.765 0.733 0.78 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
27/100 3.52G 1.252 0.6901 1.06 538 640: 100%|██████████| 18/18 [00:03<00:00, 5.03it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.08it/s]
all 70 2670 0.724 0.772 0.768 0.438
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
28/100 4.43G 1.244 0.679 1.066 373 640: 100%|██████████| 18/18 [00:03<00:00, 4.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.74it/s]
all 70 2670 0.743 0.742 0.755 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
29/100 3.64G 1.237 0.6636 1.053 387 640: 100%|██████████| 18/18 [00:03<00:00, 4.61it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.25it/s]
all 70 2670 0.756 0.771 0.786 0.442
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
30/100 4.09G 1.254 0.6769 1.056 568 640: 100%|██████████| 18/18 [00:04<00:00, 4.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.46it/s]
all 70 2670 0.767 0.772 0.795 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
31/100 3.93G 1.226 0.6508 1.052 272 640: 100%|██████████| 18/18 [00:03<00:00, 4.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.08it/s]
all 70 2670 0.784 0.758 0.821 0.446
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
32/100 4.07G 1.233 0.6573 1.068 362 640: 100%|██████████| 18/18 [00:03<00:00, 4.82it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.87it/s]
all 70 2670 0.76 0.757 0.792 0.439
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
33/100 3.91G 1.213 0.6536 1.056 374 640: 100%|██████████| 18/18 [00:03<00:00, 5.22it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.85it/s]
all 70 2670 0.78 0.774 0.809 0.454
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
34/100 4.42G 1.269 0.6727 1.061 434 640: 100%|██████████| 18/18 [00:04<00:00, 4.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.16it/s]
all 70 2670 0.782 0.753 0.797 0.452
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
35/100 4.35G 1.239 0.6634 1.046 524 640: 100%|██████████| 18/18 [00:04<00:00, 4.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.32it/s]
all 70 2670 0.752 0.76 0.801 0.447
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
36/100 4.01G 1.247 0.6559 1.059 306 640: 100%|██████████| 18/18 [00:03<00:00, 4.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.99it/s]
all 70 2670 0.763 0.734 0.763 0.426
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
37/100 3.46G 1.242 0.6598 1.063 322 640: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.17it/s]
all 70 2670 0.731 0.735 0.762 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
38/100 4.08G 1.241 0.648 1.058 347 640: 100%|██████████| 18/18 [00:03<00:00, 4.79it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.87it/s]
all 70 2670 0.724 0.728 0.749 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
39/100 4.13G 1.227 0.6396 1.055 438 640: 100%|██████████| 18/18 [00:04<00:00, 4.44it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.81it/s]
all 70 2670 0.775 0.755 0.796 0.451
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
40/100 4.33G 1.214 0.625 1.056 433 640: 100%|██████████| 18/18 [00:03<00:00, 4.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.26it/s]
all 70 2670 0.799 0.756 0.814 0.462
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
41/100 4.09G 1.205 0.6276 1.045 384 640: 100%|██████████| 18/18 [00:04<00:00, 4.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.32it/s]
all 70 2670 0.809 0.754 0.813 0.466
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
42/100 3.64G 1.227 0.6375 1.059 343 640: 100%|██████████| 18/18 [00:03<00:00, 4.91it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.49it/s]
all 70 2670 0.784 0.749 0.805 0.457
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
43/100 4.06G 1.223 0.6381 1.057 419 640: 100%|██████████| 18/18 [00:03<00:00, 4.73it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.11it/s]
all 70 2670 0.783 0.742 0.795 0.445
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
44/100 3.78G 1.22 0.6277 1.045 620 640: 100%|██████████| 18/18 [00:03<00:00, 4.85it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.59it/s]
all 70 2670 0.761 0.734 0.766 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
45/100 3.59G 1.179 0.6055 1.036 497 640: 100%|██████████| 18/18 [00:03<00:00, 5.04it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.99it/s]
all 70 2670 0.774 0.756 0.792 0.438
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
46/100 3.82G 1.21 0.6219 1.044 539 640: 100%|██████████| 18/18 [00:03<00:00, 4.88it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.85it/s]
all 70 2670 0.78 0.761 0.81 0.445
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
47/100 3.24G 1.192 0.6127 1.045 293 640: 100%|██████████| 18/18 [00:03<00:00, 4.82it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.15it/s]
all 70 2670 0.818 0.749 0.819 0.455
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
48/100 4.55G 1.176 0.6053 1.04 352 640: 100%|██████████| 18/18 [00:03<00:00, 4.64it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.69it/s]
all 70 2670 0.83 0.747 0.824 0.46
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
49/100 3.62G 1.19 0.6006 1.044 541 640: 100%|██████████| 18/18 [00:03<00:00, 4.85it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.806 0.773 0.829 0.466
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
50/100 4.24G 1.227 0.6211 1.052 407 640: 100%|██████████| 18/18 [00:03<00:00, 4.76it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.06it/s]
all 70 2670 0.815 0.754 0.811 0.454
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
51/100 3.57G 1.194 0.5965 1.05 496 640: 100%|██████████| 18/18 [00:03<00:00, 4.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.65it/s]
all 70 2670 0.812 0.778 0.826 0.458
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
52/100 4.07G 1.195 0.5946 1.037 465 640: 100%|██████████| 18/18 [00:03<00:00, 5.02it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.54it/s]
all 70 2670 0.817 0.762 0.823 0.463
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
53/100 4.12G 1.193 0.5904 1.043 553 640: 100%|██████████| 18/18 [00:03<00:00, 4.92it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.71it/s]
all 70 2670 0.785 0.784 0.82 0.466
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
54/100 4.29G 1.163 0.5937 1.041 449 640: 100%|██████████| 18/18 [00:03<00:00, 4.88it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.13it/s]
all 70 2670 0.799 0.748 0.806 0.456
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
55/100 3.79G 1.174 0.5874 1.042 262 640: 100%|██████████| 18/18 [00:03<00:00, 4.61it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.80it/s]
all 70 2670 0.785 0.761 0.791 0.444
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
56/100 3.85G 1.172 0.5826 1.03 475 640: 100%|██████████| 18/18 [00:04<00:00, 4.40it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.46it/s]
all 70 2670 0.798 0.797 0.82 0.453
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
57/100 4.27G 1.176 0.596 1.049 252 640: 100%|██████████| 18/18 [00:03<00:00, 4.72it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.66it/s]
all 70 2670 0.775 0.801 0.818 0.448
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
58/100 4.4G 1.179 0.5848 1.037 550 640: 100%|██████████| 18/18 [00:04<00:00, 4.47it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.28it/s]
all 70 2670 0.798 0.785 0.824 0.452
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
59/100 4.89G 1.167 0.5913 1.034 315 640: 100%|██████████| 18/18 [01:30<00:00, 5.03s/it]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:59<00:00, 19.85s/it]
all 70 2670 0.822 0.78 0.817 0.46
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
60/100 3.66G 1.167 0.5836 1.038 474 640: 100%|██████████| 18/18 [00:10<00:00, 1.77it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.67it/s]
all 70 2670 0.783 0.769 0.819 0.464
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
61/100 4.1G 1.169 0.5776 1.019 549 640: 100%|██████████| 18/18 [00:03<00:00, 4.72it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.97it/s]
all 70 2670 0.822 0.754 0.81 0.458
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
62/100 4.25G 1.163 0.5754 1.038 360 640: 100%|██████████| 18/18 [00:03<00:00, 4.90it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.44it/s]
all 70 2670 0.824 0.771 0.823 0.454
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
63/100 4.19G 1.164 0.5683 1.031 473 640: 100%|██████████| 18/18 [00:03<00:00, 5.05it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.81 0.735 0.799 0.444
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
64/100 4.82G 1.184 0.5701 1.031 395 640: 100%|██████████| 18/18 [00:04<00:00, 4.08it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.83it/s]
all 70 2670 0.774 0.796 0.813 0.454
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
65/100 3.62G 1.167 0.5666 1.033 511 640: 100%|██████████| 18/18 [00:03<00:00, 4.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.24it/s]
all 70 2670 0.802 0.792 0.833 0.457
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
66/100 3.68G 1.18 0.5798 1.04 411 640: 100%|██████████| 18/18 [00:04<00:00, 4.39it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.27it/s]
all 70 2670 0.838 0.784 0.831 0.463
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
67/100 3.97G 1.132 0.5516 1.034 449 640: 100%|██████████| 18/18 [00:03<00:00, 4.57it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.78it/s]
all 70 2670 0.818 0.787 0.828 0.456
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
68/100 4.84G 1.189 0.582 1.039 394 640: 100%|██████████| 18/18 [00:03<00:00, 4.53it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:05<00:00, 1.84s/it]
all 70 2670 0.821 0.775 0.825 0.457
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
69/100 3.91G 1.169 0.5716 1.034 424 640: 100%|██████████| 18/18 [00:03<00:00, 4.89it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.35it/s]
all 70 2670 0.814 0.78 0.827 0.461
EarlyStopping: Training stopped early as no improvement observed in last 20 epochs. Best results observed at epoch 49, best model saved as best.pt.
To update EarlyStopping(patience=20) pass a new patience value, i.e. `patience=300` or use `patience=0` to disable EarlyStopping.
69 epochs completed in 0.166 hours. Optimizer stripped from yolo\yolov11\train\palmeras_imgsz640_epochs100_r0\weights\last.pt, 5.5MB Optimizer stripped from yolo\yolov11\train\palmeras_imgsz640_epochs100_r0\weights\best.pt, 5.5MB Validating yolo\yolov11\train\palmeras_imgsz640_epochs100_r0\weights\best.pt... Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) YOLO11n summary (fused): 238 layers, 2,582,542 parameters, 0 gradients, 6.3 GFLOPs
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:03<00:00, 1.07s/it]
all 70 2670 0.81 0.773 0.832 0.466
Tree 43 325 0.665 0.64 0.695 0.322
Palm 66 2345 0.955 0.907 0.969 0.61
Speed: 0.4ms preprocess, 4.0ms inference, 0.0ms loss, 2.2ms postprocess per image
Results saved to yolo\yolov11\train\palmeras_imgsz640_epochs100_r0
Out[ ]:
ultralytics.utils.metrics.DetMetrics object with attributes:
ap_class_index: array([0, 1])
box: ultralytics.utils.metrics.Metric object
confusion_matrix: <ultralytics.utils.metrics.ConfusionMatrix object at 0x000001278B603450>
curves: ['Precision-Recall(B)', 'F1-Confidence(B)', 'Precision-Confidence(B)', 'Recall-Confidence(B)']
curves_results: [[array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 1, 1, 1, ..., 0.0015147, 0.00075737, 0],
[ 1, 1, 1, ..., 0.025657, 0.012828, 0]]), 'Recall', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.063038, 0.063038, 0.075066, ..., 0, 0, 0],
[ 0.33549, 0.33549, 0.41381, ..., 0, 0, 0]]), 'Confidence', 'F1'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.032593, 0.032593, 0.039065, ..., 1, 1, 1],
[ 0.20221, 0.20221, 0.26201, ..., 1, 1, 1]]), 'Confidence', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.95692, 0.95692, 0.95692, ..., 0, 0, 0],
[ 0.98422, 0.98422, 0.9838, ..., 0, 0, 0]]), 'Confidence', 'Recall']]
fitness: 0.502632982904651
keys: ['metrics/precision(B)', 'metrics/recall(B)', 'metrics/mAP50(B)', 'metrics/mAP50-95(B)']
maps: array([ 0.32175, 0.61033])
names: {0: 'Tree', 1: 'Palm'}
plot: True
results_dict: {'metrics/precision(B)': 0.8100651344029258, 'metrics/recall(B)': 0.7733291820605253, 'metrics/mAP50(B)': 0.8319734423302556, 'metrics/mAP50-95(B)': 0.46603959852402826, 'fitness': 0.502632982904651}
save_dir: WindowsPath('yolo/yolov11/train/palmeras_imgsz640_epochs100_r0')
speed: {'preprocess': 0.3856863294328962, 'inference': 4.02989387512207, 'loss': 0.0, 'postprocess': 2.171393803187779}
task: 'detect'
In [ ]:
# Se entrena el modelo base
model_yolov11_p2.train(project=RUTA_CARPETA_YOLOV11_PROYECTOS_TRAIN, **params_2)
Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) engine\trainer: task=detect, mode=train, model=yolo11n.pt, data=./resources\utils\yolov8.yaml, epochs=200, time=None, patience=20, batch=16, imgsz=480, save=True, save_period=-1, cache=False, device=cuda, workers=8, project=yolo\yolov11\train, name=palmeras_imgsz480_epochs200_r0, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=yolo\yolov11\train\palmeras_imgsz480_epochs200_r0 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 6640 ultralytics.nn.modules.block.C3k2 [32, 64, 1, False, 0.25] 3 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 4 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 5 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 6 -1 1 87040 ultralytics.nn.modules.block.C3k2 [128, 128, 1, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 346112 ultralytics.nn.modules.block.C3k2 [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 249728 ultralytics.nn.modules.block.C2PSA [256, 256, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 13 -1 1 111296 ultralytics.nn.modules.block.C3k2 [384, 128, 1, False] 14 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 15 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 16 -1 1 32096 ultralytics.nn.modules.block.C3k2 [256, 64, 1, False] 17 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 18 [-1, 13] 1 0 ultralytics.nn.modules.conv.Concat [1] 19 -1 1 86720 ultralytics.nn.modules.block.C3k2 [192, 128, 1, False] 20 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 21 [-1, 10] 1 0 ultralytics.nn.modules.conv.Concat [1] 22 -1 1 378880 ultralytics.nn.modules.block.C3k2 [384, 256, 1, True] 23 [16, 19, 22] 1 431062 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] YOLO11n summary: 319 layers, 2,590,230 parameters, 2,590,214 gradients, 6.4 GFLOPs Transferred 448/499 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir yolo\yolov11\train\palmeras_imgsz480_epochs200_r0', view at http://localhost:6006/ Freezing layer 'model.23.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks with YOLO11n... AMP: checks passed
train: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\train.cache... 279 images, 0 backgrounds, 0 corrupt: 100%|██████████| 279/279 [00:00<?, ?it/s] val: Scanning D:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\datasets\raw\palmeras_kaggle_yolo_format\test.cache... 70 images, 0 backgrounds, 0 corrupt: 100%|██████████| 70/70 [00:00<?, ?it/s]
Plotting labels to yolo\yolov11\train\palmeras_imgsz480_epochs200_r0\labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 81 weight(decay=0.0), 88 weight(decay=0.0005), 87 bias(decay=0.0) TensorBoard: model graph visualization added Image sizes 480 train, 480 val Using 8 dataloader workers Logging results to yolo\yolov11\train\palmeras_imgsz480_epochs200_r0 Starting training for 200 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
1/200 2.07G 2.46 3.72 1.711 439 480: 100%|██████████| 18/18 [00:04<00:00, 4.45it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.25it/s]
all 70 2670 0.06 0.442 0.0587 0.0289
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
2/200 2.35G 1.689 2.877 1.162 527 480: 100%|██████████| 18/18 [00:04<00:00, 4.12it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.39it/s]
all 70 2670 0.0834 0.555 0.316 0.151
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
3/200 2.3G 1.676 1.89 1.106 435 480: 100%|██████████| 18/18 [00:04<00:00, 4.40it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.54it/s]
all 70 2670 0.96 0.218 0.387 0.203
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
4/200 2.26G 1.566 1.345 1.087 262 480: 100%|██████████| 18/18 [00:03<00:00, 4.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:02<00:00, 1.42it/s]
all 70 2670 0.914 0.419 0.441 0.238
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
5/200 2.06G 1.47 1.125 1.077 265 480: 100%|██████████| 18/18 [00:03<00:00, 5.48it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.91it/s]
all 70 2670 0.823 0.378 0.381 0.2
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
6/200 2G 1.469 1.085 1.066 204 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.07it/s]
all 70 2670 0.904 0.401 0.445 0.251
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
7/200 2.08G 1.434 0.9941 1.059 408 480: 100%|██████████| 18/18 [00:04<00:00, 4.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.63it/s]
all 70 2670 0.929 0.437 0.51 0.294
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
8/200 2.8G 1.4 0.9392 1.046 477 480: 100%|██████████| 18/18 [00:03<00:00, 4.83it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.97it/s]
all 70 2670 0.742 0.519 0.596 0.326
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
9/200 2.14G 1.406 0.9041 1.062 449 480: 100%|██████████| 18/18 [00:03<00:00, 5.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.52it/s]
all 70 2670 0.734 0.451 0.547 0.31
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
10/200 2.78G 1.36 0.8727 1.046 383 480: 100%|██████████| 18/18 [00:03<00:00, 5.87it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.00it/s]
all 70 2670 0.65 0.547 0.599 0.325
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
11/200 2.36G 1.353 0.8352 1.049 426 480: 100%|██████████| 18/18 [00:02<00:00, 6.08it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.60it/s]
all 70 2670 0.596 0.621 0.619 0.334
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
12/200 2.27G 1.353 0.8165 1.042 274 480: 100%|██████████| 18/18 [00:03<00:00, 5.75it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.74it/s]
all 70 2670 0.578 0.599 0.601 0.316
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
13/200 2.26G 1.395 0.8096 1.054 531 480: 100%|██████████| 18/18 [00:02<00:00, 6.15it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:02<00:00, 1.45it/s]
all 70 2670 0.607 0.64 0.629 0.325
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
14/200 2.27G 1.39 0.7974 1.045 287 480: 100%|██████████| 18/18 [00:02<00:00, 6.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.51it/s]
all 70 2670 0.641 0.623 0.654 0.36
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
15/200 1.92G 1.385 0.811 1.053 382 480: 100%|██████████| 18/18 [00:03<00:00, 5.94it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.31it/s]
all 70 2670 0.639 0.643 0.658 0.363
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
16/200 2.42G 1.324 0.773 1.041 540 480: 100%|██████████| 18/18 [00:03<00:00, 5.74it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.06it/s]
all 70 2670 0.708 0.636 0.679 0.366
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
17/200 2.23G 1.343 0.7413 1.033 366 480: 100%|██████████| 18/18 [00:03<00:00, 5.81it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.27it/s]
all 70 2670 0.75 0.614 0.671 0.369
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
18/200 2.34G 1.328 0.7401 1.034 476 480: 100%|██████████| 18/18 [00:03<00:00, 5.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.24it/s]
all 70 2670 0.727 0.683 0.705 0.394
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
19/200 2.02G 1.342 0.7514 1.034 444 480: 100%|██████████| 18/18 [00:03<00:00, 5.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.09it/s]
all 70 2670 0.66 0.668 0.668 0.373
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
20/200 2.4G 1.307 0.7177 1.035 423 480: 100%|██████████| 18/18 [00:04<00:00, 4.44it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.70it/s]
all 70 2670 0.744 0.647 0.7 0.382
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
21/200 2.46G 1.344 0.7206 1.042 379 480: 100%|██████████| 18/18 [00:02<00:00, 6.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.20it/s]
all 70 2670 0.756 0.672 0.725 0.392
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
22/200 2.39G 1.318 0.7155 1.032 632 480: 100%|██████████| 18/18 [00:03<00:00, 5.56it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.34it/s]
all 70 2670 0.766 0.666 0.733 0.394
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
23/200 2.58G 1.314 0.7059 1.024 520 480: 100%|██████████| 18/18 [00:02<00:00, 6.35it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.29it/s]
all 70 2670 0.698 0.664 0.695 0.377
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
24/200 2.73G 1.273 0.7097 1.028 351 480: 100%|██████████| 18/18 [00:02<00:00, 6.26it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.11it/s]
all 70 2670 0.749 0.669 0.708 0.389
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
25/200 1.98G 1.28 0.7076 1.032 455 480: 100%|██████████| 18/18 [00:03<00:00, 5.82it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.18it/s]
all 70 2670 0.752 0.699 0.738 0.407
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
26/200 2.51G 1.301 0.7059 1.022 440 480: 100%|██████████| 18/18 [00:03<00:00, 5.49it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.11it/s]
all 70 2670 0.71 0.658 0.684 0.375
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
27/200 2.07G 1.296 0.7095 1.023 538 480: 100%|██████████| 18/18 [00:03<00:00, 5.55it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.86it/s]
all 70 2670 0.717 0.708 0.719 0.4
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
28/200 2.58G 1.286 0.6894 1.032 372 480: 100%|██████████| 18/18 [00:03<00:00, 5.84it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.47it/s]
all 70 2670 0.74 0.662 0.701 0.38
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
29/200 2.14G 1.277 0.6832 1.016 387 480: 100%|██████████| 18/18 [00:03<00:00, 5.66it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.02it/s]
all 70 2670 0.745 0.699 0.725 0.394
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
30/200 2.39G 1.29 0.6836 1.017 568 480: 100%|██████████| 18/18 [00:03<00:00, 6.00it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.88it/s]
all 70 2670 0.669 0.632 0.646 0.361
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
31/200 2.3G 1.269 0.6691 1.014 272 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.71it/s]
all 70 2670 0.793 0.676 0.754 0.403
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
32/200 2.38G 1.272 0.6775 1.026 363 480: 100%|██████████| 18/18 [00:02<00:00, 6.04it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.733 0.708 0.726 0.395
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
33/200 2.29G 1.249 0.6647 1.016 375 480: 100%|██████████| 18/18 [00:03<00:00, 5.07it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.33it/s]
all 70 2670 0.78 0.693 0.745 0.406
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
34/200 2.58G 1.308 0.688 1.016 434 480: 100%|██████████| 18/18 [00:03<00:00, 5.56it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.66it/s]
all 70 2670 0.816 0.699 0.761 0.414
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
35/200 2.54G 1.288 0.6739 1.012 523 480: 100%|██████████| 18/18 [00:03<00:00, 4.57it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.94it/s]
all 70 2670 0.717 0.663 0.708 0.392
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
36/200 2.34G 1.308 0.6644 1.021 306 480: 100%|██████████| 18/18 [00:04<00:00, 4.12it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.13it/s]
all 70 2670 0.727 0.68 0.708 0.39
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
37/200 2.04G 1.287 0.6648 1.024 321 480: 100%|██████████| 18/18 [00:02<00:00, 6.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.00it/s]
all 70 2670 0.69 0.693 0.714 0.397
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
38/200 2.38G 1.299 0.6641 1.023 347 480: 100%|██████████| 18/18 [00:03<00:00, 5.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.69it/s]
all 70 2670 0.748 0.719 0.756 0.416
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
39/200 2.41G 1.275 0.6614 1.022 439 480: 100%|██████████| 18/18 [00:03<00:00, 5.28it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.89it/s]
all 70 2670 0.757 0.722 0.761 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
40/200 2.53G 1.258 0.6453 1.017 433 480: 100%|██████████| 18/18 [00:03<00:00, 5.57it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.35it/s]
all 70 2670 0.741 0.713 0.755 0.415
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
41/200 2.39G 1.242 0.6385 1.009 384 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.92it/s]
all 70 2670 0.737 0.74 0.76 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
42/200 2.36G 1.273 0.6623 1.023 342 480: 100%|██████████| 18/18 [00:03<00:00, 4.76it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.61it/s]
all 70 2670 0.743 0.704 0.741 0.408
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
43/200 2.38G 1.263 0.6553 1.02 419 480: 100%|██████████| 18/18 [00:03<00:00, 5.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.77it/s]
all 70 2670 0.75 0.708 0.747 0.414
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
44/200 2.21G 1.266 0.6465 1.004 619 480: 100%|██████████| 18/18 [00:03<00:00, 5.53it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.32it/s]
all 70 2670 0.757 0.726 0.75 0.42
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
45/200 2.11G 1.237 0.6268 1.002 497 480: 100%|██████████| 18/18 [00:03<00:00, 5.41it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.63it/s]
all 70 2670 0.764 0.739 0.757 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
46/200 2.24G 1.262 0.6383 1.009 537 480: 100%|██████████| 18/18 [00:03<00:00, 5.69it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.06it/s]
all 70 2670 0.709 0.736 0.75 0.412
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
47/200 1.91G 1.25 0.6289 1.014 293 480: 100%|██████████| 18/18 [00:02<00:00, 6.01it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.90it/s]
all 70 2670 0.781 0.691 0.755 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
48/200 2.65G 1.231 0.6261 1.008 353 480: 100%|██████████| 18/18 [00:03<00:00, 5.09it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.00it/s]
all 70 2670 0.757 0.73 0.766 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
49/200 2.12G 1.236 0.6249 1.01 540 480: 100%|██████████| 18/18 [00:02<00:00, 6.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.27it/s]
all 70 2670 0.741 0.745 0.757 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
50/200 2.47G 1.273 0.641 1.015 407 480: 100%|██████████| 18/18 [00:03<00:00, 4.72it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.73it/s]
all 70 2670 0.785 0.689 0.759 0.416
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
51/200 2.1G 1.238 0.6191 1.013 495 480: 100%|██████████| 18/18 [00:02<00:00, 6.39it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.67it/s]
all 70 2670 0.747 0.755 0.771 0.424
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
52/200 2.38G 1.25 0.6211 1.003 466 480: 100%|██████████| 18/18 [00:03<00:00, 6.00it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.13it/s]
all 70 2670 0.735 0.717 0.754 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
53/200 2.41G 1.244 0.6201 1.008 553 480: 100%|██████████| 18/18 [00:03<00:00, 5.85it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.46it/s]
all 70 2670 0.782 0.729 0.768 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
54/200 2.51G 1.216 0.6192 1.006 449 480: 100%|██████████| 18/18 [00:03<00:00, 5.90it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.63it/s]
all 70 2670 0.779 0.695 0.755 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
55/200 2.22G 1.232 0.6172 1.007 261 480: 100%|██████████| 18/18 [00:03<00:00, 5.10it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.71it/s]
all 70 2670 0.779 0.73 0.766 0.423
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
56/200 2.26G 1.221 0.6075 0.9929 475 480: 100%|██████████| 18/18 [00:03<00:00, 4.55it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.38it/s]
all 70 2670 0.789 0.72 0.771 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
57/200 2.49G 1.227 0.6234 1.017 252 480: 100%|██████████| 18/18 [00:03<00:00, 5.05it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.99it/s]
all 70 2670 0.709 0.732 0.748 0.412
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
58/200 2.57G 1.229 0.6175 1.004 549 480: 100%|██████████| 18/18 [00:03<00:00, 5.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.49it/s]
all 70 2670 0.803 0.732 0.782 0.43
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
59/200 2.91G 1.226 0.6269 1.006 315 480: 100%|██████████| 18/18 [00:03<00:00, 5.29it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.27it/s]
all 70 2670 0.715 0.734 0.744 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
60/200 2.16G 1.223 0.616 1.005 475 480: 100%|██████████| 18/18 [00:03<00:00, 4.97it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.58it/s]
all 70 2670 0.699 0.687 0.723 0.405
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
61/200 2.4G 1.24 0.613 0.993 547 480: 100%|██████████| 18/18 [00:02<00:00, 6.01it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 1.94it/s]
all 70 2670 0.724 0.759 0.746 0.417
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
62/200 2.48G 1.235 0.6093 1.012 360 480: 100%|██████████| 18/18 [00:03<00:00, 5.46it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.53it/s]
all 70 2670 0.795 0.703 0.759 0.411
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
63/200 2.45G 1.228 0.602 1.003 473 480: 100%|██████████| 18/18 [00:03<00:00, 5.56it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.02it/s]
all 70 2670 0.76 0.739 0.753 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
64/200 2.8G 1.234 0.5995 0.9971 395 480: 100%|██████████| 18/18 [00:03<00:00, 5.70it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.65it/s]
all 70 2670 0.781 0.708 0.757 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
65/200 2.12G 1.212 0.6018 0.996 511 480: 100%|██████████| 18/18 [00:03<00:00, 5.42it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.28it/s]
all 70 2670 0.769 0.779 0.796 0.437
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
66/200 2.16G 1.228 0.6085 1.004 411 480: 100%|██████████| 18/18 [00:03<00:00, 5.56it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.76it/s]
all 70 2670 0.78 0.745 0.774 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
67/200 2.33G 1.189 0.5907 1.002 449 480: 100%|██████████| 18/18 [00:03<00:00, 4.97it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.28it/s]
all 70 2670 0.806 0.724 0.789 0.436
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
68/200 2.82G 1.245 0.6162 1.01 392 480: 100%|██████████| 18/18 [00:03<00:00, 5.32it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.08it/s]
all 70 2670 0.737 0.752 0.757 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
69/200 2.29G 1.228 0.6076 1.004 425 480: 100%|██████████| 18/18 [00:03<00:00, 5.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.33it/s]
all 70 2670 0.772 0.748 0.775 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
70/200 2.29G 1.199 0.5883 1.004 306 480: 100%|██████████| 18/18 [00:03<00:00, 5.16it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.92it/s]
all 70 2670 0.769 0.746 0.768 0.422
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
71/200 2.11G 1.186 0.5769 0.9964 504 480: 100%|██████████| 18/18 [00:02<00:00, 6.01it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.11it/s]
all 70 2670 0.802 0.735 0.783 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
72/200 2.51G 1.245 0.6108 1.017 321 480: 100%|██████████| 18/18 [00:04<00:00, 4.19it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.63it/s]
all 70 2670 0.755 0.738 0.77 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
73/200 2.46G 1.21 0.5892 0.9962 497 480: 100%|██████████| 18/18 [00:03<00:00, 5.87it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.17it/s]
all 70 2670 0.773 0.752 0.79 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
74/200 2.4G 1.178 0.5766 0.9848 548 480: 100%|██████████| 18/18 [00:03<00:00, 5.04it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.36it/s]
all 70 2670 0.798 0.736 0.777 0.436
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
75/200 2.16G 1.199 0.5856 0.9868 667 480: 100%|██████████| 18/18 [00:03<00:00, 5.64it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.85it/s]
all 70 2670 0.758 0.746 0.765 0.421
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
76/200 2.32G 1.214 0.5875 0.9977 537 480: 100%|██████████| 18/18 [00:03<00:00, 5.31it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.50it/s]
all 70 2670 0.777 0.746 0.77 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
77/200 2G 1.198 0.5794 0.9986 468 480: 100%|██████████| 18/18 [00:04<00:00, 4.03it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.78it/s]
all 70 2670 0.805 0.737 0.776 0.444
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
78/200 2.59G 1.197 0.578 0.9873 520 480: 100%|██████████| 18/18 [00:03<00:00, 5.96it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.52it/s]
all 70 2670 0.767 0.752 0.788 0.439
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
79/200 2.32G 1.181 0.5761 0.9916 649 480: 100%|██████████| 18/18 [00:03<00:00, 5.45it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.53it/s]
all 70 2670 0.799 0.731 0.785 0.437
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
80/200 1.95G 1.202 0.5827 1.005 302 480: 100%|██████████| 18/18 [00:03<00:00, 5.47it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.22it/s]
all 70 2670 0.798 0.7 0.755 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
81/200 2.17G 1.182 0.57 0.9914 372 480: 100%|██████████| 18/18 [00:03<00:00, 4.97it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.75it/s]
all 70 2670 0.778 0.739 0.785 0.444
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
82/200 2.75G 1.188 0.5699 1.004 436 480: 100%|██████████| 18/18 [00:03<00:00, 5.96it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.21it/s]
all 70 2670 0.784 0.752 0.776 0.439
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
83/200 2.24G 1.197 0.5812 0.9939 445 480: 100%|██████████| 18/18 [00:02<00:00, 6.06it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.62it/s]
all 70 2670 0.756 0.766 0.773 0.43
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
84/200 2.09G 1.188 0.5759 0.9929 717 480: 100%|██████████| 18/18 [00:04<00:00, 3.77it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.74it/s]
all 70 2670 0.792 0.706 0.773 0.428
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
85/200 2.08G 1.203 0.5738 0.9909 493 480: 100%|██████████| 18/18 [00:03<00:00, 4.71it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.35it/s]
all 70 2670 0.752 0.769 0.781 0.426
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
86/200 2.44G 1.195 0.5777 0.9934 366 480: 100%|██████████| 18/18 [00:02<00:00, 6.15it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.22it/s]
all 70 2670 0.791 0.751 0.783 0.436
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
87/200 2.12G 1.188 0.5754 0.9953 361 480: 100%|██████████| 18/18 [00:04<00:00, 4.24it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.33it/s]
all 70 2670 0.764 0.756 0.768 0.424
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
88/200 2.68G 1.195 0.5691 0.9957 476 480: 100%|██████████| 18/18 [00:03<00:00, 5.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.75it/s]
all 70 2670 0.764 0.763 0.791 0.437
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
89/200 2.6G 1.196 0.5738 1.002 277 480: 100%|██████████| 18/18 [00:04<00:00, 4.44it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 4.25it/s]
all 70 2670 0.777 0.782 0.799 0.441
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
90/200 2.18G 1.186 0.5644 0.992 371 480: 100%|██████████| 18/18 [00:03<00:00, 5.90it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.12it/s]
all 70 2670 0.764 0.775 0.784 0.425
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
91/200 2.43G 1.176 0.5705 0.9855 290 480: 100%|██████████| 18/18 [00:03<00:00, 5.63it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.12it/s]
all 70 2670 0.761 0.758 0.771 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
92/200 2.44G 1.173 0.5638 0.9946 346 480: 100%|██████████| 18/18 [00:03<00:00, 4.73it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.64it/s]
all 70 2670 0.79 0.746 0.781 0.438
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
93/200 2.66G 1.187 0.5646 0.9887 415 480: 100%|██████████| 18/18 [00:03<00:00, 5.67it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.74it/s]
all 70 2670 0.73 0.762 0.769 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
94/200 2.26G 1.188 0.5675 0.986 389 480: 100%|██████████| 18/18 [00:03<00:00, 5.52it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.06it/s]
all 70 2670 0.78 0.738 0.779 0.434
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
95/200 2.08G 1.171 0.5587 0.9891 494 480: 100%|██████████| 18/18 [00:03<00:00, 5.76it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.99it/s]
all 70 2670 0.739 0.75 0.76 0.419
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
96/200 2.82G 1.176 0.5728 0.9902 338 480: 100%|██████████| 18/18 [00:03<00:00, 5.80it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.48it/s]
all 70 2670 0.745 0.757 0.771 0.43
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
97/200 2.23G 1.183 0.563 0.9885 445 480: 100%|██████████| 18/18 [00:03<00:00, 4.98it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.35it/s]
all 70 2670 0.758 0.754 0.774 0.427
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
98/200 2.45G 1.161 0.5566 0.982 367 480: 100%|██████████| 18/18 [00:03<00:00, 5.83it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.08it/s]
all 70 2670 0.757 0.773 0.776 0.429
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
99/200 2.11G 1.172 0.5582 0.9854 424 480: 100%|██████████| 18/18 [00:03<00:00, 5.79it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.85it/s]
all 70 2670 0.742 0.787 0.772 0.433
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
100/200 2.41G 1.173 0.56 0.9857 310 480: 100%|██████████| 18/18 [00:03<00:00, 5.78it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:01<00:00, 2.91it/s]
all 70 2670 0.782 0.74 0.768 0.435
Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size
101/200 2.13G 1.174 0.5626 0.9951 362 480: 100%|██████████| 18/18 [00:03<00:00, 5.74it/s]
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:00<00:00, 3.36it/s]
all 70 2670 0.76 0.759 0.774 0.431
EarlyStopping: Training stopped early as no improvement observed in last 20 epochs. Best results observed at epoch 81, best model saved as best.pt.
To update EarlyStopping(patience=20) pass a new patience value, i.e. `patience=300` or use `patience=0` to disable EarlyStopping.
101 epochs completed in 0.161 hours. Optimizer stripped from yolo\yolov11\train\palmeras_imgsz480_epochs200_r0\weights\last.pt, 5.4MB Optimizer stripped from yolo\yolov11\train\palmeras_imgsz480_epochs200_r0\weights\best.pt, 5.4MB Validating yolo\yolov11\train\palmeras_imgsz480_epochs200_r0\weights\best.pt... Ultralytics 8.3.12 Python-3.11.9 torch-2.4.1+cu124 CUDA:0 (NVIDIA GeForce RTX 3060 Laptop GPU, 6144MiB) YOLO11n summary (fused): 238 layers, 2,582,542 parameters, 0 gradients, 6.3 GFLOPs
Class Images Instances Box(P R mAP50 mAP50-95): 100%|██████████| 3/3 [00:03<00:00, 1.05s/it]
all 70 2670 0.78 0.743 0.786 0.445
Tree 43 325 0.625 0.583 0.618 0.291
Palm 66 2345 0.935 0.903 0.954 0.598
Speed: 0.5ms preprocess, 2.2ms inference, 0.0ms loss, 4.8ms postprocess per image
Results saved to yolo\yolov11\train\palmeras_imgsz480_epochs200_r0
Out[ ]:
ultralytics.utils.metrics.DetMetrics object with attributes:
ap_class_index: array([0, 1])
box: ultralytics.utils.metrics.Metric object
confusion_matrix: <ultralytics.utils.metrics.ConfusionMatrix object at 0x000001DDC6E4A910>
curves: ['Precision-Recall(B)', 'F1-Confidence(B)', 'Precision-Confidence(B)', 'Recall-Confidence(B)']
curves_results: [[array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 1, 1, 1, ..., 0.00061048, 0.00030524, 0],
[ 1, 1, 1, ..., 0.014816, 0.0074078, 0]]), 'Recall', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.065083, 0.065083, 0.0857, ..., 0, 0, 0],
[ 0.33878, 0.33878, 0.45839, ..., 0, 0, 0]]), 'Confidence', 'F1'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.033777, 0.033777, 0.045027, ..., 1, 1, 1],
[ 0.20513, 0.20513, 0.29997, ..., 1, 1, 1]]), 'Confidence', 'Precision'], [array([ 0, 0.001001, 0.002002, 0.003003, 0.004004, 0.005005, 0.006006, 0.007007, 0.008008, 0.009009, 0.01001, 0.011011, 0.012012, 0.013013, 0.014014, 0.015015, 0.016016, 0.017017, 0.018018, 0.019019, 0.02002, 0.021021, 0.022022, 0.023023,
0.024024, 0.025025, 0.026026, 0.027027, 0.028028, 0.029029, 0.03003, 0.031031, 0.032032, 0.033033, 0.034034, 0.035035, 0.036036, 0.037037, 0.038038, 0.039039, 0.04004, 0.041041, 0.042042, 0.043043, 0.044044, 0.045045, 0.046046, 0.047047,
0.048048, 0.049049, 0.05005, 0.051051, 0.052052, 0.053053, 0.054054, 0.055055, 0.056056, 0.057057, 0.058058, 0.059059, 0.06006, 0.061061, 0.062062, 0.063063, 0.064064, 0.065065, 0.066066, 0.067067, 0.068068, 0.069069, 0.07007, 0.071071,
0.072072, 0.073073, 0.074074, 0.075075, 0.076076, 0.077077, 0.078078, 0.079079, 0.08008, 0.081081, 0.082082, 0.083083, 0.084084, 0.085085, 0.086086, 0.087087, 0.088088, 0.089089, 0.09009, 0.091091, 0.092092, 0.093093, 0.094094, 0.095095,
0.096096, 0.097097, 0.098098, 0.099099, 0.1001, 0.1011, 0.1021, 0.1031, 0.1041, 0.10511, 0.10611, 0.10711, 0.10811, 0.10911, 0.11011, 0.11111, 0.11211, 0.11311, 0.11411, 0.11512, 0.11612, 0.11712, 0.11812, 0.11912,
0.12012, 0.12112, 0.12212, 0.12312, 0.12412, 0.12513, 0.12613, 0.12713, 0.12813, 0.12913, 0.13013, 0.13113, 0.13213, 0.13313, 0.13413, 0.13514, 0.13614, 0.13714, 0.13814, 0.13914, 0.14014, 0.14114, 0.14214, 0.14314,
0.14414, 0.14515, 0.14615, 0.14715, 0.14815, 0.14915, 0.15015, 0.15115, 0.15215, 0.15315, 0.15415, 0.15516, 0.15616, 0.15716, 0.15816, 0.15916, 0.16016, 0.16116, 0.16216, 0.16316, 0.16416, 0.16517, 0.16617, 0.16717,
0.16817, 0.16917, 0.17017, 0.17117, 0.17217, 0.17317, 0.17417, 0.17518, 0.17618, 0.17718, 0.17818, 0.17918, 0.18018, 0.18118, 0.18218, 0.18318, 0.18418, 0.18519, 0.18619, 0.18719, 0.18819, 0.18919, 0.19019, 0.19119,
0.19219, 0.19319, 0.19419, 0.1952, 0.1962, 0.1972, 0.1982, 0.1992, 0.2002, 0.2012, 0.2022, 0.2032, 0.2042, 0.20521, 0.20621, 0.20721, 0.20821, 0.20921, 0.21021, 0.21121, 0.21221, 0.21321, 0.21421, 0.21522,
0.21622, 0.21722, 0.21822, 0.21922, 0.22022, 0.22122, 0.22222, 0.22322, 0.22422, 0.22523, 0.22623, 0.22723, 0.22823, 0.22923, 0.23023, 0.23123, 0.23223, 0.23323, 0.23423, 0.23524, 0.23624, 0.23724, 0.23824, 0.23924,
0.24024, 0.24124, 0.24224, 0.24324, 0.24424, 0.24525, 0.24625, 0.24725, 0.24825, 0.24925, 0.25025, 0.25125, 0.25225, 0.25325, 0.25425, 0.25526, 0.25626, 0.25726, 0.25826, 0.25926, 0.26026, 0.26126, 0.26226, 0.26326,
0.26426, 0.26527, 0.26627, 0.26727, 0.26827, 0.26927, 0.27027, 0.27127, 0.27227, 0.27327, 0.27427, 0.27528, 0.27628, 0.27728, 0.27828, 0.27928, 0.28028, 0.28128, 0.28228, 0.28328, 0.28428, 0.28529, 0.28629, 0.28729,
0.28829, 0.28929, 0.29029, 0.29129, 0.29229, 0.29329, 0.29429, 0.2953, 0.2963, 0.2973, 0.2983, 0.2993, 0.3003, 0.3013, 0.3023, 0.3033, 0.3043, 0.30531, 0.30631, 0.30731, 0.30831, 0.30931, 0.31031, 0.31131,
0.31231, 0.31331, 0.31431, 0.31532, 0.31632, 0.31732, 0.31832, 0.31932, 0.32032, 0.32132, 0.32232, 0.32332, 0.32432, 0.32533, 0.32633, 0.32733, 0.32833, 0.32933, 0.33033, 0.33133, 0.33233, 0.33333, 0.33433, 0.33534,
0.33634, 0.33734, 0.33834, 0.33934, 0.34034, 0.34134, 0.34234, 0.34334, 0.34434, 0.34535, 0.34635, 0.34735, 0.34835, 0.34935, 0.35035, 0.35135, 0.35235, 0.35335, 0.35435, 0.35536, 0.35636, 0.35736, 0.35836, 0.35936,
0.36036, 0.36136, 0.36236, 0.36336, 0.36436, 0.36537, 0.36637, 0.36737, 0.36837, 0.36937, 0.37037, 0.37137, 0.37237, 0.37337, 0.37437, 0.37538, 0.37638, 0.37738, 0.37838, 0.37938, 0.38038, 0.38138, 0.38238, 0.38338,
0.38438, 0.38539, 0.38639, 0.38739, 0.38839, 0.38939, 0.39039, 0.39139, 0.39239, 0.39339, 0.39439, 0.3954, 0.3964, 0.3974, 0.3984, 0.3994, 0.4004, 0.4014, 0.4024, 0.4034, 0.4044, 0.40541, 0.40641, 0.40741,
0.40841, 0.40941, 0.41041, 0.41141, 0.41241, 0.41341, 0.41441, 0.41542, 0.41642, 0.41742, 0.41842, 0.41942, 0.42042, 0.42142, 0.42242, 0.42342, 0.42442, 0.42543, 0.42643, 0.42743, 0.42843, 0.42943, 0.43043, 0.43143,
0.43243, 0.43343, 0.43443, 0.43544, 0.43644, 0.43744, 0.43844, 0.43944, 0.44044, 0.44144, 0.44244, 0.44344, 0.44444, 0.44545, 0.44645, 0.44745, 0.44845, 0.44945, 0.45045, 0.45145, 0.45245, 0.45345, 0.45445, 0.45546,
0.45646, 0.45746, 0.45846, 0.45946, 0.46046, 0.46146, 0.46246, 0.46346, 0.46446, 0.46547, 0.46647, 0.46747, 0.46847, 0.46947, 0.47047, 0.47147, 0.47247, 0.47347, 0.47447, 0.47548, 0.47648, 0.47748, 0.47848, 0.47948,
0.48048, 0.48148, 0.48248, 0.48348, 0.48448, 0.48549, 0.48649, 0.48749, 0.48849, 0.48949, 0.49049, 0.49149, 0.49249, 0.49349, 0.49449, 0.4955, 0.4965, 0.4975, 0.4985, 0.4995, 0.5005, 0.5015, 0.5025, 0.5035,
0.5045, 0.50551, 0.50651, 0.50751, 0.50851, 0.50951, 0.51051, 0.51151, 0.51251, 0.51351, 0.51451, 0.51552, 0.51652, 0.51752, 0.51852, 0.51952, 0.52052, 0.52152, 0.52252, 0.52352, 0.52452, 0.52553, 0.52653, 0.52753,
0.52853, 0.52953, 0.53053, 0.53153, 0.53253, 0.53353, 0.53453, 0.53554, 0.53654, 0.53754, 0.53854, 0.53954, 0.54054, 0.54154, 0.54254, 0.54354, 0.54454, 0.54555, 0.54655, 0.54755, 0.54855, 0.54955, 0.55055, 0.55155,
0.55255, 0.55355, 0.55455, 0.55556, 0.55656, 0.55756, 0.55856, 0.55956, 0.56056, 0.56156, 0.56256, 0.56356, 0.56456, 0.56557, 0.56657, 0.56757, 0.56857, 0.56957, 0.57057, 0.57157, 0.57257, 0.57357, 0.57457, 0.57558,
0.57658, 0.57758, 0.57858, 0.57958, 0.58058, 0.58158, 0.58258, 0.58358, 0.58458, 0.58559, 0.58659, 0.58759, 0.58859, 0.58959, 0.59059, 0.59159, 0.59259, 0.59359, 0.59459, 0.5956, 0.5966, 0.5976, 0.5986, 0.5996,
0.6006, 0.6016, 0.6026, 0.6036, 0.6046, 0.60561, 0.60661, 0.60761, 0.60861, 0.60961, 0.61061, 0.61161, 0.61261, 0.61361, 0.61461, 0.61562, 0.61662, 0.61762, 0.61862, 0.61962, 0.62062, 0.62162, 0.62262, 0.62362,
0.62462, 0.62563, 0.62663, 0.62763, 0.62863, 0.62963, 0.63063, 0.63163, 0.63263, 0.63363, 0.63463, 0.63564, 0.63664, 0.63764, 0.63864, 0.63964, 0.64064, 0.64164, 0.64264, 0.64364, 0.64464, 0.64565, 0.64665, 0.64765,
0.64865, 0.64965, 0.65065, 0.65165, 0.65265, 0.65365, 0.65465, 0.65566, 0.65666, 0.65766, 0.65866, 0.65966, 0.66066, 0.66166, 0.66266, 0.66366, 0.66466, 0.66567, 0.66667, 0.66767, 0.66867, 0.66967, 0.67067, 0.67167,
0.67267, 0.67367, 0.67467, 0.67568, 0.67668, 0.67768, 0.67868, 0.67968, 0.68068, 0.68168, 0.68268, 0.68368, 0.68468, 0.68569, 0.68669, 0.68769, 0.68869, 0.68969, 0.69069, 0.69169, 0.69269, 0.69369, 0.69469, 0.6957,
0.6967, 0.6977, 0.6987, 0.6997, 0.7007, 0.7017, 0.7027, 0.7037, 0.7047, 0.70571, 0.70671, 0.70771, 0.70871, 0.70971, 0.71071, 0.71171, 0.71271, 0.71371, 0.71471, 0.71572, 0.71672, 0.71772, 0.71872, 0.71972,
0.72072, 0.72172, 0.72272, 0.72372, 0.72472, 0.72573, 0.72673, 0.72773, 0.72873, 0.72973, 0.73073, 0.73173, 0.73273, 0.73373, 0.73473, 0.73574, 0.73674, 0.73774, 0.73874, 0.73974, 0.74074, 0.74174, 0.74274, 0.74374,
0.74474, 0.74575, 0.74675, 0.74775, 0.74875, 0.74975, 0.75075, 0.75175, 0.75275, 0.75375, 0.75475, 0.75576, 0.75676, 0.75776, 0.75876, 0.75976, 0.76076, 0.76176, 0.76276, 0.76376, 0.76476, 0.76577, 0.76677, 0.76777,
0.76877, 0.76977, 0.77077, 0.77177, 0.77277, 0.77377, 0.77477, 0.77578, 0.77678, 0.77778, 0.77878, 0.77978, 0.78078, 0.78178, 0.78278, 0.78378, 0.78478, 0.78579, 0.78679, 0.78779, 0.78879, 0.78979, 0.79079, 0.79179,
0.79279, 0.79379, 0.79479, 0.7958, 0.7968, 0.7978, 0.7988, 0.7998, 0.8008, 0.8018, 0.8028, 0.8038, 0.8048, 0.80581, 0.80681, 0.80781, 0.80881, 0.80981, 0.81081, 0.81181, 0.81281, 0.81381, 0.81481, 0.81582,
0.81682, 0.81782, 0.81882, 0.81982, 0.82082, 0.82182, 0.82282, 0.82382, 0.82482, 0.82583, 0.82683, 0.82783, 0.82883, 0.82983, 0.83083, 0.83183, 0.83283, 0.83383, 0.83483, 0.83584, 0.83684, 0.83784, 0.83884, 0.83984,
0.84084, 0.84184, 0.84284, 0.84384, 0.84484, 0.84585, 0.84685, 0.84785, 0.84885, 0.84985, 0.85085, 0.85185, 0.85285, 0.85385, 0.85485, 0.85586, 0.85686, 0.85786, 0.85886, 0.85986, 0.86086, 0.86186, 0.86286, 0.86386,
0.86486, 0.86587, 0.86687, 0.86787, 0.86887, 0.86987, 0.87087, 0.87187, 0.87287, 0.87387, 0.87487, 0.87588, 0.87688, 0.87788, 0.87888, 0.87988, 0.88088, 0.88188, 0.88288, 0.88388, 0.88488, 0.88589, 0.88689, 0.88789,
0.88889, 0.88989, 0.89089, 0.89189, 0.89289, 0.89389, 0.89489, 0.8959, 0.8969, 0.8979, 0.8989, 0.8999, 0.9009, 0.9019, 0.9029, 0.9039, 0.9049, 0.90591, 0.90691, 0.90791, 0.90891, 0.90991, 0.91091, 0.91191,
0.91291, 0.91391, 0.91491, 0.91592, 0.91692, 0.91792, 0.91892, 0.91992, 0.92092, 0.92192, 0.92292, 0.92392, 0.92492, 0.92593, 0.92693, 0.92793, 0.92893, 0.92993, 0.93093, 0.93193, 0.93293, 0.93393, 0.93493, 0.93594,
0.93694, 0.93794, 0.93894, 0.93994, 0.94094, 0.94194, 0.94294, 0.94394, 0.94494, 0.94595, 0.94695, 0.94795, 0.94895, 0.94995, 0.95095, 0.95195, 0.95295, 0.95395, 0.95495, 0.95596, 0.95696, 0.95796, 0.95896, 0.95996,
0.96096, 0.96196, 0.96296, 0.96396, 0.96496, 0.96597, 0.96697, 0.96797, 0.96897, 0.96997, 0.97097, 0.97197, 0.97297, 0.97397, 0.97497, 0.97598, 0.97698, 0.97798, 0.97898, 0.97998, 0.98098, 0.98198, 0.98298, 0.98398,
0.98498, 0.98599, 0.98699, 0.98799, 0.98899, 0.98999, 0.99099, 0.99199, 0.99299, 0.99399, 0.99499, 0.996, 0.997, 0.998, 0.999, 1]), array([[ 0.88923, 0.88923, 0.88615, ..., 0, 0, 0],
[ 0.97228, 0.97228, 0.97143, ..., 0, 0, 0]]), 'Confidence', 'Recall']]
fitness: 0.4788239075167768
keys: ['metrics/precision(B)', 'metrics/recall(B)', 'metrics/mAP50(B)', 'metrics/mAP50-95(B)']
maps: array([ 0.29108, 0.59824])
names: {0: 'Tree', 1: 'Palm'}
plot: True
results_dict: {'metrics/precision(B)': 0.7800100413229583, 'metrics/recall(B)': 0.7431596755850487, 'metrics/mAP50(B)': 0.7863008086461691, 'metrics/mAP50-95(B)': 0.44465980739128874, 'fitness': 0.4788239075167768}
save_dir: WindowsPath('yolo/yolov11/train/palmeras_imgsz480_epochs200_r0')
speed: {'preprocess': 0.48579829079764225, 'inference': 2.2145680018833707, 'loss': 0.0, 'postprocess': 4.814679282052176}
task: 'detect'
Verificación de Resultados¶
Como parte de los resultados de entrenamiento, YOLO también valida los mismos, de acuerdo con lo descrito en el archivo de configuración (.yaml). A continuación una comparación de las pérdidas en entrenamiento vs. validación, y la precisión mAP50 (accuracy con un IoU de 0.5).
In [ ]:
# Se grafican los resultados de los modelos YOLOv8
ruta_resultados_csv_1 = os.path.join(RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}_r0', 'results.csv')
ruta_resultados_csv_2 = os.path.join(RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_2}_epochs{YOLO_EPOCHS_2}_r0', 'results.csv')
ruta_resultados_v11_csv_1 = os.path.join(RUTA_CARPETA_YOLOV11_PROYECTOS_TRAIN, f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}_r0', 'results.csv')
resultados_df_1 = pd.read_csv(ruta_resultados_csv_1)
resultados_df_2 = pd.read_csv(ruta_resultados_csv_2)
resultados_df_v11_1 = pd.read_csv(ruta_resultados_v11_csv_1)
# Se grafican los resultados para las pérdidas en entrenamiento y validación
plt.figure(figsize=(12, 6))
plt.plot(resultados_df_1['epoch'], resultados_df_1['train/box_loss'], label=f'Pérdida en Entrenamiento - YOLOv8_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}', color='blue')
plt.plot(resultados_df_1['epoch'], resultados_df_1['val/box_loss'], label=f'Pérdida en Validación - YOLOv8_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}', linestyle='--', color='blue')
plt.plot(resultados_df_2['epoch'], resultados_df_2['train/box_loss'], label=f'Pérdida en Entrenamiento - YOLOv8imgsz{YOLO_IMG_SIZE_2}_epochs{YOLO_EPOCHS_2}', color='green')
plt.plot(resultados_df_2['epoch'], resultados_df_2['val/box_loss'], label=f'Pérdida en Validación - YOLOv8imgsz{YOLO_IMG_SIZE_2}_epochs{YOLO_EPOCHS_2}', linestyle='--', color='green')
plt.xlabel('Epoch')
plt.ylabel('Loss')
plt.title('Périda en Entrenamiento vs Validación para los modelos YOLOv8')
plt.legend()
plt.grid(True)
plt.show()
# Se grafican los resultados para el mAP50
plt.figure(figsize=(12, 6))
plt.plot(resultados_df_1['epoch'], resultados_df_1['metrics/mAP50(B)'], label=f'mAP50 - YOLOv8_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}', color='blue')
plt.plot(resultados_df_2['epoch'], resultados_df_2['metrics/mAP50(B)'], label=f'mAP50 - YOLOv8_imgsz{YOLO_IMG_SIZE_2}_epochs{YOLO_EPOCHS_2}', color='green')
plt.xlabel('Epoch')
plt.ylabel('mAP50')
plt.title('mAP50 para los modelos YOLOv8')
plt.legend()
plt.grid(True)
plt.show()
# Se grafinca una comparación de mAP50 para los modelos YOLOv8 y YOLOv11
plt.figure(figsize=(12, 6))
plt.plot(resultados_df_1['epoch'], resultados_df_1['metrics/mAP50(B)'], label=f'mAP50 - YOLOv8_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}', color='blue')
plt.plot(resultados_df_v11_1['epoch'], resultados_df_v11_1['metrics/mAP50(B)'], label=f'mAP50 - YOLOv11_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}', color='red')
plt.xlabel('Epoch')
plt.ylabel('mAP50')
plt.title('mAP50 para los modelos YOLOv8 y YOLOv11')
plt.legend()
plt.grid(True)
plt.show()
Se puede observar que una imagen de mayor tamaño ofrece mejores resultados en términos de mAP50. Sin embargo, esto trae un costo computacional mayor; por lo que depende del hardware del que se disponga. En cuanto a la precisión del modelo v8 con el más reciente v11, no se nota mucha diferencia. Sin embargo, la mejora del modelo puede encontrarse en las versiones de mayor tamaño (y complejidad).
Inferencia¶
In [20]:
RUTA_CARPETA_YOLOV8_BEST_TRAINED = os.path.join(RUTA_CARPETA_YOLOV8_PROYECTOS_TRAIN, f'{NOMBRE_MODELO_YOLO}_imgsz{YOLO_IMG_SIZE_1}_epochs{YOLO_EPOCHS_1}_r0', 'weights', 'best.pt')
RUTA_CARPETA_YOLOV8_BASE = 'yolov8n.pt'
# Se cargan los mejores pesos del modelo entrenado
inf_model_yolo_v8 = YOLO(RUTA_CARPETA_YOLOV8_BEST_TRAINED)
inf_model_yolov8n = YOLO(RUTA_CARPETA_YOLOV8_BASE)
IMAGENES_PARA_INFERENCIA_PATH = os.path.join(RUTA_CARPETA_RESOURCES, 'imgs', 'inferencia')
RESULTADOS_INFERENCIA_PATH = os.path.join(IMAGENES_PARA_INFERENCIA_PATH, 'yolo_results')
# Se lanzan las inferencias
#set_imagenes = ['mvd_001.jpg']
results = inf_model_yolo_v8(IMAGENES_PARA_INFERENCIA_PATH, classes=[1], save=True, project=RESULTADOS_INFERENCIA_PATH, name='yolov8_retrained')
results_base = inf_model_yolov8n(IMAGENES_PARA_INFERENCIA_PATH, save=True, project=RESULTADOS_INFERENCIA_PATH, name='yolov8_base') # Clases por defecto
# Se procesan los resultados
for result in results:
boxes = result.boxes # Objeto Boxes para las salidas de detección
masks = result.masks # Objet Mask para las salidas de segmentación
keypoints = result.keypoints # Objeto Keypoints para las salidas de posición
probs = result.probs # Objeto Probs para las salidas de clasificación
obb = result.obb # Objeto OBB (oriented bounding box)
result.show()
for result in results_base:
boxes = result.boxes # Objeto Boxes para las salidas de detección
masks = result.masks # Objet Mask para las salidas de segmentación
keypoints = result.keypoints # Objeto Keypoints para las salidas de posición
probs = result.probs # Objeto Probs para las salidas de clasificación
obb = result.obb # Objeto OBB (oriented bounding box)
result.show()
image 1/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_001.jpg: 544x640 40 Palms, 8.6ms image 2/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_002.jpg: 416x640 32 Palms, 9.4ms image 3/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_003.jpg: 480x640 39 Palms, 9.3ms Speed: 3.2ms preprocess, 9.1ms inference, 1.3ms postprocess per image at shape (1, 3, 480, 640) Results saved to resources\imgs\inferencia\yolo_results\yolov8_retrained image 1/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_001.jpg: 544x640 (no detections), 8.2ms image 2/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_002.jpg: 416x640 1 potted plant, 11.0ms image 3/3 d:\ProyectosUBA\CEIA\5_Vision_por_Computadora_II\repo_bruno\uba-ceia\ceia-vpc2\resources\imgs\inferencia\inf_003.jpg: 480x640 (no detections), 9.0ms Speed: 2.0ms preprocess, 9.4ms inference, 1.7ms postprocess per image at shape (1, 3, 480, 640) Results saved to resources\imgs\inferencia\yolo_results\yolov8_base
🔮 Función auxiliar:
Función que recibe una lista de imágenes en formato de tensores y las grafica en una fila.
In [38]:
def show_images_in_line_from_tensors(tensor_images: List[torch.tensor]) -> None:
"""
Función que recibe una lista de imágenes en formato de tensores y las grafica en una fila.
Parámetros:
- tensor_images: Lista de tensores que representan las imágenes (deben ser exactamente 3 tensores).
"""
if len(tensor_images) != 3:
raise ValueError("La función requiere exactamente tres tensores de imágenes.")
# Convertir a formato PIL para visualización con Matplotlib
images = [T.ToPILImage()(img) for img in tensor_images]
# Crear la figura con tres subplots
fig, axes = plt.subplots(1, 3, figsize=(15, 5))
# Mostrar cada imagen en su subplot correspondiente
for i, img in enumerate(images):
axes[i].imshow(img)
axes[i].axis('off') # Ocultar los ejes
# Mostrar las imágenes
plt.show()
Para realizar una validación previa del modelo, se utilizan imágenes obtenidas de servicios satelitales que nunca fueron vistas por el mismo. Las imágenes bases son:
In [39]:
base_dir = os.path.join(RUTA_CARPETA_RESOURCES, 'imgs', 'inferencia')
img_base_list = [read_image(os.path.join(base_dir, img_path)) for img_path in os.listdir(base_dir) if img_path.endswith(FORMATOS_IMAGENES)]
show_images_in_line_from_tensors(img_base_list)
Predicciones del modelo YOLOv8 base (sin fine-tuning):
In [40]:
base_dir = os.path.join(RUTA_CARPETA_RESOURCES, 'imgs', 'inferencia', 'yolo_results', 'yolov8_base')
img_base_list = [read_image(os.path.join(base_dir, img_path)) for img_path in os.listdir(base_dir) if img_path.endswith(FORMATOS_IMAGENES)]
show_images_in_line_from_tensors(img_base_list)
Predicciones del modelo YOLOv8 con fine-tuning:
In [41]:
base_dir = os.path.join(RUTA_CARPETA_RESOURCES, 'imgs', 'inferencia', 'yolo_results', 'yolov8_retrained')
img_base_list = [read_image(os.path.join(base_dir, img_path)) for img_path in os.listdir(base_dir) if img_path.endswith(FORMATOS_IMAGENES)]
show_images_in_line_from_tensors(img_base_list)
Conclusiones¶
⚠️ Problemas detectados:
pycocotools, lo que resulta un problema para adaptarla a un código estandar.pytorch, especialmente para las rotaciones y cambios de perspectivas, no se aplicaban correctamente a las bounding-boxes. Se tuvo una mejora con Albumentations, pero igualmente no quedaron perfectos. Un ejemplo de esto es la siguiente imagen:
Figura 5: Data-aumentation fallida.
⭐ Conclusión ⭐
Como conclusiones, se pudo observar que el tamaño de la imagen es claramente un hiperparámetro que está limitado por los recursos. También la rigidez de los modelos pre-entrenados, así como su facilidad para utilizarlos out-of-the-box. En este sentido, se valora mucho frameworks como el de utltralitycs que es muy fácil su implementación en un dataset propio. Otro punto importante es el consumo de memoria de estos modelos, en donde es importante, para casos de producción, estudiar una estrategia de entrenamiento distribuído, y también hacer las aumentaciones offline, dado que el modelo sin aumentaciones demoró mucho menos su entrenamiento. Como conclusión final, se observa que hay varias matices que resolver, pero sin duda, estos enfoques sirven para aplicaciones en producción, destacando modelos como YOLO, debido a su alta performance y principalmente fácil implementación.
💫 Mejoras posibles:
torchscript o onnx para ver su tamaño.prediction=True
Referencias¶
- Adel Ammar, and Anis Koubaa. (2023). Aerial images of palm trees [Data set]. Kaggle. https://doi.org/10.34740/KAGGLE/DSV/6382990
- TorchVision Object Detection Finetuning Tutorial — PyTorch Tutorials 2.4.0+cu121 documentation. (n.d.). https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html
Anexo¶
Si hay problemas de GPU, probar con las siguientes soluciones:
!pip install GPUtil
import torch
from GPUtil import showUtilization as gpu_usage
from numba import cuda
def free_gpu_cache():
print("Initial GPU Usage")
gpu_usage()
torch.cuda.empty_cache()
cuda.select_device(0)
cuda.close()
cuda.select_device(0)
print("GPU Usage after emptying the cache")
gpu_usage()
free_gpu_cache()
import gc def report_gpu(): print(torch.cuda.list_gpu_processes()) gc.collect() torch.cuda.empty_cache()
# En el entrenamiento:
def train_loop(files, batch_size, model):
size = len(files)
for batch in range(size//batch_size):
im_paths = files[batch*batch_size : (batch+1)*batch_size]
X,y = load_batch(im_paths)
# Compute prediction and loss
output = model(X, y)
del X
del y
gc.collect()
torch.cuda.empty_cache()
# Chequear donde está el modelo
next(model.parameters()).device
# Chequear donde está un tensor
tensor.device